The Nineteenth Byte - 2021-09-17 (page 2 of 4)

« first day (3884 days earlier) ← previous day next day → last day (952 days later) »

pVC aecidiospore adduced

7:16 AM

heres a list of some basic elements i am gonna implement for dinoux

CMQ: ^^ Thoughts?

7:36 AM

@pVCaecidiosporeadduced Some APL influence…

@pVCaecidiosporeadduced What is the difference between ₂ and ²?

pVC aecidiospore adduced

@Adám one is 2^x other is x²

> Raise to the power of 2

Should probably say "Raise 2 to the power of"

pVC aecidiospore adduced

fixed

> ≠ Lesser than or equal to, vectorised.

pVC aecidiospore adduced

what; fixed

7:50 AM

@pVCaecidiosporeadduced Seems a lot of operations are missing, e.g. for sorting descending (̀S or ś?), e or e-to-the-power-of, logarithms, factors, OR, 0…n-1 (ɹ?), etc.

pVC aecidiospore adduced

ah those ones are diagraphs; havent put them yet

is sort descending actually useful?

sort then reverse

Right, more useful for grading.

Are you not worried about confusing bitwise xor with increment? ^ vs ⌃

pVC aecidiospore adduced

hmm i will change them

i need to find those tiny versions of > and <

@pVCaecidiosporeadduced Why not use ⊻ for xor?

pVC aecidiospore adduced

actually i orginally wanted to use the tiny (or squished) versions of > and <

7:56 AM

For increment/decrement?

pVC aecidiospore adduced

yes

I'd have a hard time remembering which one would be up and which one down.

How about using harpoons or other arrows?

pVC aecidiospore adduced

a > increment as in number line and < decrement

↾ and ⇂ or ↥ and ↧ or ⤒ and ⤓ or ⭱ and ⭳ or ⟑ and ⟇ or ⩓ and ⩔ etc.

pVC aecidiospore adduced

ok i used the first 2

8:03 AM

:-) You're welcome.

8:16 AM

user image

8:38 AM

Yay, I have 15k rep! I can... protect questions... yay.

3

pVC aecidiospore adduced

9:16 AM

@Adám added some of those

to the list

Vyxal uses ‹› for inc/dec

9:54 AM

Oh, Puzzling.

10:05 AM

@pxeger Hmm...

2 hours later…

pVC aecidiospore adduced

11:46 AM

@emanresuA thank you

11:59 AM

@pVCaecidiosporeadduced No you know about shapecatcher.com – best tool for language design.

caird coinheringaahing

12:18 PM

@pVCaecidiosporeadduced A very useful trick: . is short for 0.5 in Jelly, and due to the indexing behaviour, .ị gets the 0th and 1st elements of a list, i.e the first and last elements

Jelly tip about using that

12:53 PM

0

A: Sandbox for Proposed Challenges

Apply gravity to this matrix Given an \$m\times n\$ matrix of \$0\$'s and \$1\$'s, apply "gravity" to it. This means to drop down all the \$1\$'s as if they were affected by gravity. For example [[1,0,1,1,0,1,0] [0,0,0,1,0,0,0] [1,0,1,1,1,1,1] [0,1,1,0,1,1,0] [1,1,0,1,0,0,1]] Should result i...

caird coinheringaahing

1:21 PM

Oooh nice, I have exactly 42000 rep on main :P

Now to get 7 upvotes and to downvote an answer :P

pVC aecidiospore adduced

@Adám cool

so why are some elements repeated in langs?

@cairdcoinheringaahing check your rep

Browncat Programs

@pVCaecidiosporeadduced No it's not

Your data is wrong

pVC aecidiospore adduced

@BrowncatPrograms ok

Browncat Programs

; is the most common atom, € is the most common symbol

pVC aecidiospore adduced

@cairdcoinheringaahing its super close

@BrowncatPrograms concatenate?

Browncat Programs

1:28 PM

Yeah

caird coinheringaahing

@pVCaecidiosporeadduced If you undo the downvote, I can downvote something to get to 42069 :P

Browncat Programs

caird and I both have independently developed corpuses (corpii?), which mostly agree on the top 10

caird coinheringaahing

Also, not entirely sure I'm happy about that voting :/

pVC aecidiospore adduced

@cairdcoinheringaahing ok

Browncat Programs

I think a skunk died in my high school

Fr

pVC aecidiospore adduced

1:30 PM

@cairdcoinheringaahing now you have to lose a single rep

@pVCaecidiosporeadduced Was this question for me? And if so, can you explain what you mean in more detail?

caird coinheringaahing

user image

4

Behold: funny numbers :P

pVC aecidiospore adduced

@Adám no

caird coinheringaahing

@pVCaecidiosporeadduced status-completed :P

@cairdcoinheringaahing haha funny number go brrrrrrr

pVC aecidiospore adduced

1:37 PM

@Adám like in jelly there is U and Ṛ

both say they are reverse?

caird coinheringaahing

U reverses at depth 1, Ṛ is flat

pVC aecidiospore adduced

@cairdcoinheringaahing so x[::-1]

caird coinheringaahing

Try them both with [[1,2,3],[4,5,6]] as input

pVC aecidiospore adduced

@cairdcoinheringaahing i dont know jelly

caird coinheringaahing

Hang on, I'll get a TIO for you

1:41 PM

@pVCaecidiosporeadduced U reverses each item while R underdot just does the whole list

caird coinheringaahing

@pVCaecidiosporeadduced Try it online!

Top line is result of U

@lyxal Yes but actually no

pVC aecidiospore adduced

oh

so now i would have to implement these

@cairdcoinheringaahing how no?

caird coinheringaahing

@lyxal U reverses at depth 1, so it recursively looks for lists of numbers, and reverses those lists. For example, Try it online!

The second line would be "reverse each"

0

Q: Setting up Al's R session

Al wanted to use my computer to do some simple calculations, so I lent him the use of my R terminal for a bit. He complained though, saying it wasn't set up how he liked it. I said "No problem Al, set it up how you like and I'll just revert it later." This suited Al, and I left him to it. After a...

programming-puzzle r

pVC aecidiospore adduced

1:48 PM

i should probably just learn a golfing lang rather than making my own :/

Browncat Programs

No, making your own is a good idea

It'll give you a good understanding of how they work

caird coinheringaahing

Both are good ideas tbh

@NewPosts Feels weird to review this as "Looks OK", but I'm not sure what else I'd do :/

pVC aecidiospore adduced

So is CGCC getting any IGS recently?

Browncat Programs

No

Just generic spam

pVC aecidiospore adduced

@BrowncatPrograms i think i already know how they work (or maybe not)

Browncat Programs

1:56 PM

An even better understanding of how they work then :p

I made my first golfing language (attempt), Ash, before I tried learning any golfing languages

@emanresuA ~~no comment~~

pVC aecidiospore adduced

@BrowncatPrograms how did ash go?

how many Pokémons did he get? :P

Browncat Programs

I decided to stop working on it as it was consuming too much of my time and the parsing code was getting too scary

Plus I'd made some stupid decisions early on, that were holding the language back

pVC aecidiospore adduced

@BrowncatPrograms what were those?

this is the 2nd rewrite of dinoux

Browncat Programs

Some decisions about the code page and its layout, the general structure of the language, and some operator choices

pVC aecidiospore adduced

2:00 PM

@BrowncatPrograms codepage?

howdo you implement it?

Browncat Programs

Wdym?

Code page as in SBCS

pVC aecidiospore adduced

isnt it placing 256 chars on a table and calling it sbcs

Browncat Programs

I mean sort of yeah

But you want to make sure the characters are aesthetically pleasing, correctly spaced/rendered, nicely organized, easy to differentiate and associate with the right operator, grouped by function, etc.

pVC aecidiospore adduced

@BrowncatPrograms how does that make a difference?

Browncat Programs

It doesn't make it golfier, but it makes the language nicer to use

pVC aecidiospore adduced

2:03 PM

?

how?

Browncat Programs

Well, if operators that look similar have similar functions (i.e., lower cased letters always being dyads), it makes it easier to read and write the language

pVC aecidiospore adduced

also should the comment char be in the codepage?

Browncat Programs

You probably shouldn't waste a code page space on comments

pVC aecidiospore adduced

@BrowncatPrograms like jelly

Browncat Programs

Yep

@BrowncatPrograms But if you put it outside the code page, it means comments can never be used in an answer using the SBCS

2:06 PM

golfing languages do not need comments

@pVCaecidiosporeadduced How about ignoring everything after an unrecognised unquoted character and until the end of the line?

because they don't need to be readable

Browncat Programs

You might want to try Pip's method of two spaces being a comment

at all

pVC aecidiospore adduced

@BrowncatPrograms vyxal

Browncat Programs

2:06 PM

@Razetime I disagree

They should be as readable as possible until it gets in the way of golfiness

and since your language is stack based you can just push a string to stack and immediately delete it

makeshift comments

pVC aecidiospore adduced

@Adám no i am using apl's comments

@Razetime hm....i dont have strings

@pVCaecidiosporeadduced So why not leave that char out of the charset?

no string syntax?

pVC aecidiospore adduced

i am trying to make dinoux readable

@Adám ok

@Razetime ???

2:08 PM

This gives you comments for free.

again, you're making a golflang, readability is the last thing you need to be worried about unless that's a key goal like in vyxal

pVC aecidiospore adduced

Strings are useless.

Sep 14 at 5:56, 35 minutes total – 47 messages, 4 users, 0 stars

Bookmarked Sep 14 at 6:34 by pVC aecidiospore adduced

Browncat Programs

Vyxal's the opposite of readable lol

Jelly's readable, Vyxal isn't

lies

Browncat Programs

Readability is very important for a golfing language imo

pVC aecidiospore adduced

2:09 PM

nothing is readable if you dont know the lang :P

Browncat Programs

They're still used by humans

On a good day i can't tell the difference between jelly and vyxal code at a glance

Browncat Programs

So they should be human usable

There's a lot of ways to accomplish that that don't compromise golfiness

Like having a good code page

Having used it for a while now, I find Vyxal to be extremely readable/usable.

husk is the most readable golflang

Browncat Programs

2:10 PM

Husk is pretty readable, yeah

Jelly and Husk are my two favorite golflangs for aesthetics/readability/code page design

pVC aecidiospore adduced

H Return the head of the array. Ḥ Return everything except the head. Ḣ Pop the head. T Return the tail of the array. Ṭ Return everything except the tail. Ṫ Pop the tail.

^ head and tail are H and T

@BrowncatPrograms ehh ok

most golflangs look awful in default fonts due to codepage

all the explanations are misaligned

Browncat Programs

Exactly

best looking golflang is APL /s

Browncat Programs

The characters should be chosen to ensure they're all the same width in most fonts

pVC aecidiospore adduced

2:12 PM

@Razetime use fixed sys, the best font

Browncat Programs

@BrowncatPrograms That's a huge thing for readability

pVC aecidiospore adduced

use iosevka with terminal spacing

you're forced to use latin1 if you want a "readable" golflang, and you end up putting 5 different versions of 'A' that are barely distinguishable from each other, combined with othe randoms symbols you're forced to assign functions to

Jul 17 at 2:42, by Aaron Miller

Solution: Now we just need to make Vyxal Mono font

golflangs are not meant to be readable.

programmable, sure but not readable

Browncat Programs

2:14 PM

@Razetime Uh...no?

There're plenty of characters in various blocks that render fine in most fonts

@AaronMiller mod an existing font maybe

Browncat Programs

Jelly is almost never misaligned, but it has characters from all over the place

Plus its code page has a really cool feature

Every character on it can be typed on a US-INTL keyboard

That's the coolest part about it

how do you type them

alt codes?

@AaronMiller Fairfax HD has the entire Vyxal codepage

@BrowncatPrograms Having characters that are easy to type has nothing to do with readability, though. I'd much rather have ⌈ and ⌊ for ceiling and floor than Ċ and Ḟ. The symbol makes it much more easy to see what the command does than a simple letter, even if it may throw of the alignment of an explanation. The readability of the program is increased by using those characters.

Browncat Programs

⌈ renders at the wrong width in most fonts, though

(And by "readability" I also mean "writeability")

2:21 PM

i doubt that any jelly golfer types jelly on their keyboard

it's either the language bar or copypaste

Browncat Programs

I do whenever I use Jelly

@BrowncatPrograms I'd argue that they are two very different things, personally.

Browncat Programs

I don't see why y'all are opposed to readability though

I could see not caring, but how could readability ever be a negative thing?

On the contrary, readability is an integral part of Vyxal.

Browncat Programs

Maybe in theory :p

2:23 PM

it's almost as if two people can have different opinions of what's readable or not surprised pikachu face

Browncat Programs

lol

@BrowncatPrograms do you use jelly as a scripting language

@rak1507 What? Impossible! /s

⌈⌊ over ĊḞ any day though

Browncat Programs

@Razetime No, but it's still a language humans use

So it should be usable for humans

And if that comes at no cost to the byte count, why not?

caird coinheringaahing

2:24 PM

@Razetime If I could figure out how to config my second keyboard to do so, I would :P

Browncat Programs

@rak1507 I have to disagree there :p

I always thought being unreadable and hard to use is part of the fun of golflangs(or even esolangs in general).

@BrowncatPrograms fair enough

Browncat Programs

I think we just have differing priorities on what exactly makes a golfing language usable

caird coinheringaahing

Golflangs aren't unreadable tho, so long as you know what the builtins do

Browncat Programs

2:25 PM

@cairdcoinheringaahing Unless it's Ash

I regularly use Vyxal for things like formatting text, converting to/from ascii codes, generating large repeating texts, etc. that are unrelated to code golfing.

caird coinheringaahing

Same way that Java is unreadable if you don't know what the various builtins do

I know what all the Vyxal builtins do, they solve FizzBuzz!

5

caird coinheringaahing

Seriously, if you show me a piece of code from any lang that I don't know, esoteric or practical, I'll tell you it's unreadable, because duh

It's like if someone were to show you some text in Swahili or Tamil or something, you'd say it's unreadable.

caird coinheringaahing

2:27 PM

@AaronMiller 256 bytes to solve FizzBuzz? I thought Vyxal was supposed to have a short FizzBuzz program :P

Browncat Programs

I will also offer a +100 bounty to whoever posts the 30kth starred message

@cairdcoinheringaahing Readability > Golfability tho

Browncat Programs

We have hit 29900 in the last few hours

caird coinheringaahing

@AaronMiller Exactly. "Unreadable" is a stupid way to describe a language, because of course it's unreadable if you don't know it

Browncat Programs

Well, "unreadable" could mean "the learning curve to read it is too steep", which is a perfectly fine criticism

2:28 PM

Heck, I know English and I'd still say it's unreadable sometimes! :P

@BrowncatPrograms I'd argue that Jelly has a steeper learning curve than Vyxal, but I suppose that's not the case for everyone.

Browncat Programs

For the language itself, definitely

caird coinheringaahing

Without a doubt it does

Browncat Programs

Stack based is way easier than Jelly's tacit

But when you factor out the necessary complexity, I think Jelly's more elegant and intuitive

But anyway, gtg o/

caird coinheringaahing

I'd argue that Jelly's style of infix tacitness (meaning it relies on specific parsing rules) is the most difficult form to understand

pVC aecidiospore adduced

@BrowncatPrograms well my messages get starred rarely

2:31 PM

I learned Vyxal and got the bounty for 5 answers in a single day. It just made sense to me. I've tried learning Jelly twice and can't seem to get the hang of it.

caird coinheringaahing

Husk's prefix tacit means that you can more or less parse it mentally without having to know any rules aside from "how many arguments does this function take?"

pVC aecidiospore adduced

and i cant learn any golflang

caird coinheringaahing

APL's tacitness can get weird, with the arity being determined by the pattern, but it's fairly easy to get started (and difficult to master)

@pVCaecidiosporeadduced well apparently you can program in your own golflang

caird coinheringaahing

With Jelly, it makes sense until you start adding in dyads, at which point you need to understand the parsing rules, which can be fairly complex

2:34 PM

Also, with things like ‘ or ’ instead of › or ‹, I generally find Vyxal's commands to make more sense for what they do.

does jelly repeat arguments?

caird coinheringaahing

What do you mean?

a tacit system has to plug in arguments

caird coinheringaahing

@AaronMiller Even Dennis has gone on record saying he regrets choosing ‘’ :P

so imagine I plug in a,b,c and there's an empty space to fill

does it plug a again

caird coinheringaahing

2:35 PM

@Razetime Oh right. Yeah, if you have a dyadic chain with 1 argument, it reuses the left argument

@AaronMiller Most of Jelly's commands, being letter based, have mnemonics to remember them by, if you can learn them :P

For example, an underdot often (but not always) means that it is the inverse of the dotless function (H is halve, Ḥ is unhalve (double)). Most of the commands' mnemonics are words that start with the letter (İnverse, ḃijective base, żip)

husk doesn't auto repeat, I suppose that helps with some of the reading

Browncat Programs

3:13 PM

@AaronMiller That's the program layout not things I'm talking about, like code page readability

I'm mostly talking about how well you can tell what each character is and what it does, not necessarily how the program is structured

Since Jelly having a much more complex structure is a feature, not a bug

So I noticed on this question that there are several answers that are technically invalid because they use the wrong output format. I've commented on a couple where I know the posters are still active here, but there are also some older ones where the posters might not be around anymore. Should I comment there too? Flag for moderator attention? Leave well enough alone?

caird coinheringaahing

I'd say leave a comment, then come back to them tomorrow/in a few days, and flag for mod attention if unaddressed

@BrowncatPrograms Corpora. It's a third-declension neuter noun. :P

@cairdcoinheringaahing Makes sense, thanks

3:45 PM

Among other intresting things, this contains some ideas that could be nice for golfing challenges, in case someone is interested...

files.righto.com/calculator/sinclair_scientific_simulator.html

4:00 PM

1 hour ago, by Aaron Miller

Also, with things like ‘ or ’ instead of › or ‹, I generally find Vyxal's commands to make more sense for what they do.

Browncat Programs

1 hour ago, by caird coinheringaahing

@AaronMiller Even Dennis has gone on record saying he regrets choosing ‘’ :P

> I generally find Vyxal's commands to make more sense for what they do

@BrowncatPrograms Quick explanation for someone who doesn't Jelly?

Browncat Programs

‘’ is a very cherry picked example though

@DLosc Those are increment and decrement

Ah, okay

My biggest frustration in putting together a SBCS has been the number of really nice mathematical symbols that don't monospace in supposedly monospace fonts >_<

Browncat Programs

4:05 PM

Same :/

@BrowncatPrograms More examples include ; for concatenation, or _ for subtraction, although again, that's just my thoughts.

Browncat Programs

Ash had those exact operators for those two things and I hadn't even seen Jelly's operator list before

I think they're pretty intuitive

But + and - are even more intuitive to me

isn't + addition

also i suppose you could use - for subtraction but then you'd need like _ or ¯ or smth else for negatives

+ overloads to concatenation for strings. It's intuitive because it's addition for strings.

4:13 PM

@AaronMiller [laughs in Julia]

Also, why does Jelly have an identity command?

I assume it makes some tacit thing easier or something

@AaronMiller so then how do you concatenate lists

you Join them

my main complaint with vyxal is that half the time there's a good array builtin i need it's only available for strings because it's an overload of a vectorizing math operator

@AaronMiller purely for chaining

Browncat Programs

Strings and arrays should behave identically with almost all operators, IMO

4:16 PM

jelly's approach to that is to simply not have strings :P

caird coinheringaahing

@hyper-neutrino Also sometimes for filtering

@DLosc For example

granted, and I think I've said this before, jelly is pretty bad for string manipulation most of the time and vyxal is considerably better (not sure how it compares to 05AB1E which is also better than jelly for strings), but if there's a math/array problem, i'm sticking to jelly

That's understandable. Every language has its strengths and weaknesses

@DLosc That's decidedly different from how Vyxal multiplies strings: For example

is that behavior actually useful

caird coinheringaahing

4:19 PM

@hyper-neutrino That said, Jelly's string support would be much better with some more dedicated string manipulation atoms

true

caird coinheringaahing

It;s string weakness isn't in the typing of strings, it's in how the language deals with them

i suppose jelly could do string overloading by just observing when the argument is a list of characters, since those usually don't work with any mathematical operations

@hyper-neutrino Well I have used it before :P

Browncat Programs

@hyper-neutrino That feels kind of hacky though

4:21 PM

One of the nice things about the overloading in Vyxal is that it basically just goes "If this errors because of typing, make an overload," which means that there are pretty much no errors due to typing

there are no errors due to typing (as in argument types) but there are errors due to typing (as in code typos or improper implementation) :P

4

lol there are some exceptions, though: Try it Online!

@hyper-neutrino Yeah, the subtraction/negation/negative numbers trio can be problematic for golflang codepage design. Pip gets away with using - for all three, but only because of the non-tacit infix syntax, which is less golfy in general.

toki pona and code golf fans unite! toki sona, a language with 14 tokens and a 1000-character interpreter: github.com/dmadisetti/sona.js

@AaronMiller lol I just realized, the beginning of the code for that cop looks like somebody's reaction after seeing the flags: "oh f***"

4:26 PM

> why
why not

Perfect ^_^

Browncat Programs

4:37 PM

@AaronMiller That's a pretty common thing to do

oh lol I've only ever really looked at Jelly so I figured it wasn't typical

05AB1E doesn't seem to overload anything either

Browncat Programs

They're both like 5 years old

It's pretty well known now that errors are typically bad

caird coinheringaahing

@hyper-neutrino jli (Jelly v2) was intended to have only 2 types: lists and numbers, and numbers would have an internal flag to indicate character-ness for printing

Essentially, every operation on numbers would be as well-defined on characters as they are on numbers

oh that's interesting

i had that as an idea too but it felt too restrictive to not have strings at all

Browncat Programs

@cairdcoinheringaahing That sounds like a really bad idea

caird coinheringaahing

4:48 PM

Dennis also played with character arithmetic in Jelly before essentially marking the idea as status-deferred :P

Browncat Programs

@BrowncatPrograms It takes away any opportunities for overloading for literally no reason

caird coinheringaahing

But it doesn't

does being able to do math operations on characters even really offer much?

Browncat Programs

@cairdcoinheringaahing Well then what's the difference between "character" and "number with character flag"? A fancy name?

caird coinheringaahing

Overloading doesn't rely on the "types", instead it simply modifies the internal flag

4:49 PM

does vyxal have a flag that automatically converts I/O between characters and codepage index / codepoint

Are you thinking of K?

~~idk how to program in K~~ yeah looks like it

Try it Online!

caird coinheringaahing

@hyper-neutrino Kind of

If you want to get the next character, it's a lot easier to increment than it is to convert to numbers, increment, convert back

@AaronMiller Try it Online!

caird coinheringaahing

4:53 PM

@BrowncatPrograms I saw some prototypes that Dennis was playing with a few years ago, and they seemed to save bytes more than add them

Browncat Programs

For what sort of challeges? All I can imagine that being helpful for is character arithmetic

Unrelated String

jelly doesn't exactly have a ton of overloading so i can't see there being a loss there

caird coinheringaahing

@BrowncatPrograms No idea, I'd have to go digging through the Jelly room transcripts

Browncat Programs

To me "numbers with an internal character flag" sound like the exact same thing as characters without overloading

Unrelated String

still cleaner than length-1 python strings

caird coinheringaahing

4:56 PM

What's wrong with ḤL being the same as L on strings? :P

Browncat Programs

@UnrelatedString Sounds like an implementation detail?

Unrelated String

i'm barely following this conversation because i'm in the middle of a semantics lecture but isn't this half about implementation details

and then you get to have completely illegal character values in intermediate steps of computation

Browncat Programs

Sounds like an extraordinarily niche need

Another idea, please critique: There is no character type; strings are a separate type that is essentially a list of numbers, where the list is automatically flattened and the numbers are automatically converted to ints modulo codepage size.

Browncat Programs

Sounds like a string type with extra steps :p

« first day (3884 days earlier) ← previous day next day → last day (952 days later) »

Transcript for

Sep '2117

The Nineteenth Byte

The Nineteenth Byte: General discussion for codegolf.stackexc...

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile