@dzaima no I'm not arguing what the definition of "BACONS NOTHING" is, I'm arguing the meaning of the word, and that it is confusing and a different name should be chosen... and this isn't bike shedding, its conceptually inside-out and hurts my brain

you can't have a nothing that is something nothing is the absence of "thinginess"

because something is joined to x

@nathanrogers It isn't really entirely omitted in cases where it is speical either - e.g. {(1+𝕨)-𝕩}4 isn't (1+)-4

it is a thing, and thus cannot be "nothing"

@nathanrogers and thus {𝕨‿𝕩}5 errors, as it must be a thing, but cannot

how about no errors and just x

what is 1 + nothing?

@nathanrogers it's nothing

nothing, right? so I would assume it returns -x

because nothing -x... well, that's just -x

so then why is nothing‿x an exception to that rule?

@nathanrogers if anything, it'd have to be ‿𝕩 which is stupid

right, and since there is nothing to join, there IS NOTHING TO JOIN, so its just x

right do you see how this confuses and contorts the words?

nothing is an inappropriate word to use in the case where nothing can be something

@nathanrogers ‿ isn't joining, it's creating an array out of items. 𝕨∾𝕩 would be join

the two phrases are parallel

what about 1‿2‿𝕨‿3‿4 - should that be 1‿2‿3‿4?

The unfortunate fact: Programmers have to accept what the language designer calls things, even if some things sound nonsense to them.

no this doesn't sound nonsense, it is mental torture

unless, what I'm proposing is adopted

(also, the fact that the length of a‿b‿c might not be 3 is absolutely enough for me to instantly not accept your definition)

@dzaima what would you do if one of those things were nothing? It wouldnt' exist in the list!

Like, an empty list is not nothing

just as an empty bucket is not nothing

nothing is to be without even a bucket

there is no bucket, so there is NO THING

Your bucket analogy failed many times already

right, how is this hard

wtf?

@nathanrogers I'd say the list wouldn't exist. a‿b‿c must be 3-item list by the syntax, but if one is a nothing, you can't have 3 items, therefore error

no that's literally what it is.
Empty list ←→ empty bucket
List       ←→ bucket with things
Nothing    ←→ NO FUCKING THING ALREADY

And please stop using f-word

fold? :)

Words mean things. "nothing" is a word which has meaning

what would {𝕨‿𝕨}1 be in your mind?

nothing

It wouldn't evaluate

because the function does nothing and returns nothing

the only tokens defined in the scope represent nothing, and so nothing should happen

@nathanrogers ok that's just strange. 1‿1‿𝕨‿𝕨 is a 2-item array. 1‿𝕨‿w is a scalar atom 1. 𝕨‿𝕨 is.. an error? that's INCREDIBLY inconsistent

but so long as you choose to use nothing

@nathanrogers y'know, you don't have to take literally every word in the most absolutely literal meaning

no I'm just using the intuitive meaning of what nothing is

and in the terms of the notation

it doesn't elevate thought

it creates cognitive dissonance

@nathanrogers well you have weird intuition

no

nothing is no thing

unexistence

as i said, with the literal interpretation, {(1+𝕨)-𝕩}5 would be (1+)-5 but you agree that that's wrong

right

so choose a new word

null, empty, void

even nada

I'm ok with nada

@nathanrogers empty can be confused for an empty array. Null can be confused with the classical null pointer (not that BQN has those). void doesn't really describe what it is, and doesn't match how other langs use it either. nothing at least has clear resemblance

idiom: "Nothing from nine leaves nine."

nothing - 9 is not 0 - 9

its -9

@dzaima the behavior doesn't reflect the name chosen

What is the exact behavior in the first place?

@nathanrogers It reflects it much better than any other proposal you have

either the behavior should change, or the name should change

plain and simple

@nathanrogers there you're clearly taking the absolute literal meaning of "nothing"

@nathanrogers because this is just madness

@dzaima nothing really doesn't have a lot of nuance

it means no thing. not anything. not a single thing, without properties of thinginess

that's what nothing means

its literally in the word

NO

... wait for it

THING

why do my links not work?

@nathanrogers In BQN, it means there is no value for the current expression (the expression's value is nothing). It doesn't mean a token that's absolutely ignored from the syntactical representation of the code.

How about "placeholder"?

@nathanrogers "There is no value before x in the 2-item array". That is indeed nonsensical, therefore error. Is that so unclear?

right

that's because

if there is no value in the array before x

then it is just X, and not an array

or its a 1 element array, not 2

I just read how it works as a right arg (any function applied returns self immediately) and left arg (the function ignores it) and everywhere else (errors). This does sound more than "nothing"

and IMO it's doing even more job than what a "null" or "void" can represent

so my proposal of "placeholder"

@Bubbler yeah, placeholder is okay. Still think "nothing" is fine though.

call it NIL or ZILCH

the obvious name to call it is

No, I don't think it should be named as a word that implies zero or null-ness

maybenothing?

@dzaima (though placeholder sounds like it could be "filled" or something)

call it maybenothing

or just maybe

maybe is 5 chars longer than

@ngn you have a point

Just thought about something funny: it looks like a Pac-man facing left, as it consumes everything on its left then dies

@nathanrogers imo it's pretty clear that ·‿1 read as "an array of nothing and 1" is clearly implying that there are 2 items, the first is a nothing, and the second is a 1. But an array can't have nothings as items

what does @Bubbler?

@dzaima... this is also an array of nothing and one [1]

there's NOTHING on either side

@nathanrogers I disagree

it isn't a question of what " ·" is in your language dude

its a matter of this is how people speak

@Bubbler No, the programmers really don't have to use my terminology. I control how I talk about the language, not how anyone else does.

I come into a bacon forums asking questions, and I say "like hey guys, for some reason, I'm adding this value, but there's nothing happening, nothing is being appended to my list!"

like that's retarded

@nathanrogers [1] is not an array of nothing AND 1 DEFINITELY

when I can't say things that make sense

@nathanrogers it was all for

@dzaima you're arguing implementation details, I'm arguing :57350413 the language is perverse

@nathanrogers So using synonyms for "nothing" definitely doesn't fix the problem, to the extent there is one.

@Marshall that's why I suggested maybe

@nathanrogers That isn't possible though.

its a maybe value

it might be a value

it might not

@nathanrogers no, I'm arguing about language there. I believe [1] is not an array of "nothing" and "1". It might be an array which contains "nothing" and "1", but not an array which fully consists of precisely "nothing and 1"

[maybe, 1] is just [1]

Well, you can always use · as the "name" of the object in question, and let people call it what they want

@dzaima if I append nothing to the list [1], what do I get?

[1]

obviously, you get

@nathanrogers but you don't get [1,nothing], which you'd need to if 1‿𝕨 were to be 1

1,nothing is 1

there is nothing there

@Marshall why isn't that possible?

a maybe value is simultaneously a value and control flow

@Bubbler In fact I nearly always write Nothing (·) to avoid confusion. In spoken BQN I would probably find another way to disambiguate.

Let's just call it Middle Dot

@Marshall it isn't a matter of disambiguating the bacon, its a matter of disambiguating for general purpose conversation

@Marshall I like to explicitly say "a nothing"

so call it a_nothing, which is also cognitive dissonance because you can't have 1 of... NO THING

@dzaima i have against that

@nathanrogers Nothing isn't a value, so it can't be passed as a value. It can't be assigned as a name or used as the last expression in a non-namespace block.

@Marshall I get that, I read the tutorial, but its still not nothing, its maybe

maybe something maybe nothing

but if its nothing, then it IS NO THING

1,maybe is either 1,value or 1

should be a value?

not 1,value 1, ??????

@nathanrogers that's what 𝕨 is. Whether it's a nothing depends on the runtime situation

@dzaima no because maybe is valid to append to a list

maybe,1 when maybe is nothing is just 1

@nathanrogers You're describing 𝕨, not ·. These are the only two tokens (okay, 𝕎 too) that can ever indicate Nothing; · always indicates it and 𝕨 sometimes does.

en.wikibooks.org/wiki/Haskell/Understanding_monads/Maybe

@nathanrogers right, 𝕨 is what you call a "maybe", but · is a "nothing"

but w is the only thing that can ever BE a nothing

therefore w is ACTUALLY a maybe

@nathanrogers right. so?

and then nothing can be nothing

@nathanrogers 𝕨 is a thing that can either be a regular value, or a nothing. If you want to call that a "maybe", go ahead, whatever

right but the nothing is just an implementation detail at this point

that means.... NOTHING

and else matters

what is important is that W is a maybe

THAT means something

so just conceal the nothing, and just say W is a maybe

then ∧/W_X makes sense regardless of the context

well, there is · which is always a nothing

but it can't be used ever

you can't assign it to something, you can't assign to it

you can't pas sit as an argument

@nathanrogers it can, it trains most importantly

what

sure, it won't last long, but that doesn't matter

show me something where that makes sense

@nathanrogers (÷·-1+⊢) is the reciprocal of the negation of one plus argument

why not ∘

@nathanrogers because then you might need parentheses in more complex situations which is just ugly

@Bubbler which is what I've been saying

Left one I know, the problem is the right one

@Bubbler that seems wrong. dzaima/BQN gives ¯4‿¯4

Oh. Time to raise an issue then

@dzaima (÷∘-1+⊢)

•←(÷∘-1+⊢)4 works just fine

@dzaima That's a cool way to use monadic functions in trains, which looks like J's [: g h but better

@Bubbler yeah, that's precisely what it is

trains suck

@Marshall @dzaima reading through the tutorial, it seems to me that the only context in which "nothing" is relevant is when discussing whether a function is monadic or dyadic... in that sense, conceal the "nothing" as an implementation detail, and say that W is a "maybe"

@nathanrogers how about (÷·+´1+⊢)?

idk, what about it

@nathanrogers transforming it with ∘ makes it an ugly (÷∘(+´)1+⊢)

I thought bacon uses modifers to avoid this sort of thing

And the code is broken if you remove the inner parens

i thought modifiers bind more tightly than functions

Modifiers bind from left to right

So ÷∘+´ is (÷∘+)´

and, next step up, (1+·+´1+⊢) - writing that without · would be either (1+(+´1+⊢)) or (1+⟜(+´)1+⊢)

herpa derp

Try (+´1+⊢)4‿5 instead

right because you can't sum an atom?

@nathanrogers yeah. ´ requires that the argument is a vector and nothing else

•←÷(+´ 1 + ⊢)4‿5

still

@nathanrogers F←÷(+´ 1+ ⊢) vs F←÷ ·+´ 1+ ⊢ though

You mean the result is the same, or BQN is not working?

yeah that's getting dumb

(÷∘+´1+⊢)4‿5‿6 and (÷·+´1+⊢)4‿5‿6 give different results

I'm all for eliminating parens where possible... I'm not for creating loads of miscommunication potential and cognitive dissonance

@Bubbler notice I didn't send that

@nathanrogers You didn't send what?

they return different results because like you said ÷ binds to + not +'

Yes.

yeah, that's why I wrote it as an atop instead

OK then

@dzaima I'm getting a parse error for your solution anyway

nvm

or is that an over

I never remember

over

@nathanrogers the bigger one is the one that does more. so ∘ executes both functions just once, whereas ○ will execute its right operand twice for dyadic application