Ash

Feeds

2:19 AM

Lyxal has added Redwolf Programs to the list of this room's owners.

Lyxal

Welcome y'all

Not my golflang but redwolfs

Feeds

Lyxal has removed Lyxal from the list of this room's owners.

Lyxal

Okay so there's a few things I want to discuss here

1. Why do you feel you need a take n but bad?

Challenges usually allow you to choose 0 or 1 indexing

So have 1 or the other

2. Halving on lists should reduce the number of elements, not add more

Like maybe remove first and last

Instead of double each element

3

+-*/ should probably do something useful with two non-number strings

And how does one convert a list to a number?

Link to specs

(pin that somebody ^)

@RedwolfPrograms

Redwolf Programs

2:36 AM

@Lyxal That's not the final spec; I have and will be making changes to it

1. I think one indexing can be useful sometimes for some math stuff, but I guess I get your point. I will still keep the Range 1 to N for sure, and maybe the For But Bad (those were temporary names of course)
2. That makes sense, will do
3. I want certain operations to always cast to a certain type, like arithmetic, concatenation, and map/filter

Lyxal

@RedwolfPrograms you know you can edit gists right?

Redwolf Programs

I'm starting on the final spec for the wiki, so I don't really see a point. At the moment the gist is mostly up to date.

Oh here's my repo for it, there's only one page in the wiki atm

5

It's all still sort of a mess, give me a few days and the docs and interpreter should be good

Working on the string compression right now

Unrelated String

3:49 AM

just a heads up about the code page, ⋘ is just barely narrower than two properly monospaced characters, so code explanations could get reeeeeeeal ugly

Redwolf Programs

@UnrelatedString That's what tabs are for, right?

Shouldn't be too annoying with an auto-generator

Lyxal

@UnrelatedString I can vouch for that: Keg has quite a few extra wide chars in its code page. Explanations got really unaligned and annoying

@RedwolfPrograms no, they screw up the spacing

Redwolf Programs

Wait really?

How?

Lyxal

They make those tabs very marginally misaligned

Unrelated String

I remember that one time I used Neim and that HURT

Redwolf Programs

3:54 AM

@Lyxal I don't notice any misalignment

Unrelated String

any character that isn't the normal monospace width in DVSM bust be purged

Lyxal

evidence a

Unrelated String

@RedwolfPrograms tio.run/##S03OSylITkzMKcop@P//UfeMGq6ampr//wE

Redwolf Programs

Oh wait, the editor handles them differently from the result :/

Unrelated String

yeah it's a font thing

Lyxal

3:54 AM

evidence b

Notice how there is a very slight misalignment with the pounds

You see this happen with charcoal

The comments in explanations get real messy

Redwolf Programs

Who the heck makes tab characters not go to the proper alignments, kinda defeats the purpose of them :/

Lyxal

Which is why you'd want to have a more paragraph based explanation

@RedwolfPrograms those characters we're telling you not to use

They do it

Redwolf Programs

Yeah but tabs should still be designed to line up regardless of what's before them, right?

That's the point of them. Even typewriters do that :/

Unrelated String

wait isn't most of charcoal's codepage funky fullwidth
that does actually align with itself

Lyxal

Not if they have a longer width

Unrelated String

3:57 AM

@RedwolfPrograms regardless you misalign what comes afterwards

Lyxal

@UnrelatedString I tried making an explanation once it was real messy

Unrelated String

f

Lyxal

@RedwolfPrograms ^^^

Redwolf Programs

I've already got a bunch of weird width characters, and I like them, so it seems easier to just keep them

Unrelated String

it even messes up the column widths of your code page table lmao

⊶ would be cool if it monospaced though

good taste

Redwolf Programs

4:18 AM

Okay, I can't decide if I should use the term function, operator, or command

On the one hand I use terms like arity and argument that would suggest they're functions, and the way they're treated is sort of like functions (but not in the functional programming sense)

On the other hand, calling them functions might make people think they have features like first class functions and optional arguments, which isn't possible

Unrelated String

I don't think people would necessarily assume that

Redwolf Programs

So you think it's fine to call them functions?

Unrelated String

I'd definitely not go for operator unless they're infix

yeah

maybe I'm biased because mine are also called functions

Redwolf Programs

Postfix is the correct term for when it's operator operand*, right?

Unrelated String

yeah

you could also say suffix but postfix is better

main reason I'm saying functions is I'm ripping off APL terminology

despite not knowing APL

Razetime

4:45 AM

so, this one's stack based as well?

Redwolf Programs

No, not stack based

Razetime

interesting

non-stack based postfix

I'll definitely be trying this

Lyxal

5:13 AM

@RedwolfPrograms atoms

call them atoms

atoms can have arity

and atoms can have arguments

@RedwolfPrograms

Redwolf Programs

I like function more, atom would definitely work though

Redwolf Programs

5:50 AM

room topic changed to Ash: Chat for the Ash golfing language by Redwolf Programs (no tags)

Lyxal

fl=2

Is that prime checking?

Because it's length(factors(input)) == 2

which checks to see if the list of factors is [1, input]

Redwolf Programs

It'd be =lf2

Lyxal

so prefix?

Redwolf Programs

Oh, I guess that's prefix not postfix

Honestly I'm not really sure which I prefer

Lyxal

Wait how does it know to group lf together

how does it know I don't want f2

Redwolf Programs

5:55 AM

Oh, you'd have to put something after f as input, so =2lf would probably be better

Lyxal

Okay but how does it know it's not =2l then f.

Redwolf Programs

Implicit input isn't really going to be easily possible with the system I'm using, which is unfortunate

Well it knows l is unary and 2 is nullary

Lyxal

Ah I see what you mean about implicit input being difficult

Because the parser would see = (arity 2) and say "I need two things". It'd collect the 2` and then collect the l and see that the l wasn't complete

but then what happens if I have =2l

without the f

Redwolf Programs

My current design would use implicit input in that case

Lyxal

ġF1 10

prints 1\n2\n3...\n10

Redwolf Programs

6:01 AM

Pretty much, except the ` 10` would be replaced with a constant of some sort

Looks like I didn't actually remember to add a constant for ten though so that might be a good thing to do

Lyxal

İ"Heck""e""*"

Censors the word heck, making it h*ck

Redwolf Programs

I'm planning on having “ be able to end the string when it encounters something outside one of a few character classes, so you'd be able to golf that to İ“Heck“e“*

Lyxal

@RedwolfPrograms I literally just had that ready to send

i got ninja'd

@RedwolfPrograms Tastes like Jelly

Maybe you could have different arity patterns that expand in different ways

throw a bit of tacitness into the mix

Redwolf Programs

I'm going to have four types of string literals: single character, single UTF-8 character, the one above, and typical half-compressed strings

Lyxal

@RedwolfPrograms 'e '∞ "Heck" "İĿ"?

Redwolf Programs

6:12 AM

Pretty much, but using those ugly smart quote characters instead

Lyxal

stupid smart quotes

Redwolf Programs

So for the tacitness, are you saying something like unary functions being treated the same way, but binary operators are infix instead?

o5+o4=o9 instead of =+o5o4o9

Lyxal

6:31 AM

No tacit isn't necessarily infix

Redwolf Programs

Well I know it's a lot more complicated than that but I'm not just going to make it entirely tacit or it becomes Jelly with different builtins :p

Lyxal

Hang on I'm trynna remember what the chain rules of jelly are

Unrelated String

they depend on arity

Lyxal

Do you have a plan for parsing yet

Unrelated String

and basically the only thing that isn't left to right is if a monadic chain has a dyad then a monad or a dyadic chain has two dyads

aside from nilad dyad pairs obviously

Redwolf Programs

6:33 AM

@Lyxal Mostly, yes

Lyxal

@RedwolfPrograms I felt that ping

2x-1

10:26 AM

@Lyxal Use 2 indexing so that you can have a built-in returning the second-to-last item of a list.

@Lyxal Polymorphism can be less useful than automatic type casting.

Lyxal

@2x-1 how so?

Unrelated String

10:46 AM

I feel like autocasting might be best considered as a kind of polymorphism

where you more so want polymorphic behavior that's conceptually similar to casting than you want strict casting everywhere

like, look at how Jelly builtins that handle lists can't decide if they want to turn integers into singleton lists, ranges, or digit lists

it just works

Redwolf Programs

10:35 PM

Parsing string literals is not going to be fun. I'm going to change it so that they'll be parsed when necessary (i.e., when the function they're passed to is called) instead of during tokenization.

Lyxal

What?

Why?

Redwolf Programs

Two reasons:
1. It's boring and I want to do the fun part right now
2. The real reason, which is that the string compression allows some of the shorthand strings to be given custom values, and I'd rather that be possible midway through the program

Lyxal

11:07 PM

@RedwolfPrograms so you mean uncompressing then

Redwolf Programs

I guess yeah. I don't know much interpreter/compiler terminology

Lyxal

Tokenization -> Parsing -> Code Generation

Redwolf Programs

I have made it way simpler though by removing any UTF-8 stuff and using Unicode code points directly

Lyxal

That's the three steps

Tokenisation is where you give everything a name

Redwolf Programs

I know that much I just don't know much about what parsing does because I've never gotten that far :p

Lyxal

11:09 PM

You say "this string of digits is a number"

And "this here bracket is syntax"

Parsing is grouping these tokens into constructs or structures

You say "this bracket and this number followed by this block start and next 3 tokens is a structure"

And "these tokens between one token and another is another construct"

Redwolf Programs

That's useful, thanks!

Lyxal

Code generation is where you actually do the execution

You go through each construct (for lack of better word) and evaluate it

Trust me I know this...my software course at school from 2001 taught us the stages of interpreting/compiling

Redwolf Programs

I just came up with something kind of cool

Lyxal

@RedwolfPrograms take a look at Vyxal's parser. You might get some ideas

Redwolf Programs

I'm making ’ use the following byte as a code point between 80 and ff, or if it's lower than that it takes the byte after it as well so you can compactly as possible represent any character from 0080 to 7fff using that one operator