The Nineteenth Byte

2EZ 4RTZ

00:19

@trichoplax Oh I wasn't saying it was quite

I was just the only one.

it was kinda nice

trichoplax

The chat can sometimes be quiet for a few hours, but as soon as someone says something there's someone around to respond, because they were getting on with something else with the tab in the background

fəˈnɛtɪk

Does anyone know what the euler transform that is being referred to by A335 in OEIS is?

HyperNeutrino

00:35

@2EZ4RTZ *quiet

@fəˈnɛtɪk There are at least three types ಠ_ಠ

2EZ 4RTZ

:_()

fəˈnɛtɪk

So it appears that the one referred to is probably the second one there

HyperNeutrino

oh is it the cumulative sum of A292?

huh not quite

fəˈnɛtɪk

It could be the third though

yeah. looks like it is supposed to be the third...

Conor O'Brien

01:10

@miles I'm trying to find a nice way of writing the first adverb in this TIO link in J. Is there some nice adverb I'm missing? I feel like J has something like this, but my memory is failing me. Simple concatenate-then-fold doesn't work, nor any "clever" idea for a non-trivial verb.

2EZ 4RTZ

01:20

Soo I believe they made an AI that beats pro Dota 2 players

it learned over several lifetimes for humans how to play

DJMcMayhem

@ConorO'Brien What's an adverb? (Explained for someone who knows a little Jelly, and zero J)

HyperNeutrino

A quick?

Conor O'Brien

@DJMcMayhem do you know the each thingy in jelly? the currency sign? that's an adverb basically

HyperNeutrino

I think verb/adverb means atom/quick.

Conor O'Brien

adverb is basically a function that operates on verbs

HyperNeutrino

01:29

(so we're essentially saying the same thing xD)

like a meta-function for function compositions

The terminology for Ceres is the same as J for verb/adverb :P

DJMcMayhem

@ConorO'Brien Oh, like €?

Conor O'Brien

well, except jelly doesn't have nouns in the code (IIRC), so the meaning of verb/adverb differs slightly

@DJMcMayhem that's the one

HyperNeutrino

@ConorO'Brien Wait what is a noun then o0

DJMcMayhem

Ok. Is reduce also an adverb?

Conor O'Brien

IIRC jelly implements data as constant functions?

@DJMcMayhem yup

HyperNeutrino

01:31

@ConorO'Brien Yes. Niladic expressions.

(or nilad atoms)

Conor O'Brien

yeah

those are Jelly's closest thing to nouns

a noun is a literal piece of data, like 3 or 234

adverbs and conjunctions operate on both verbs and nouns

(a conjunction is a dyadic adverb)

DJMcMayhem

Oh. That's weird

HyperNeutrino

Ah okay. Interesting

DJMcMayhem

Do you have an example of a conjunction?

HyperNeutrino

join

like $

Conor O'Brien

01:32

in jelly? something like the if thingy

does jelly have "compose"?

DJMcMayhem

Or J, if you explain it simply

HyperNeutrino

@ConorO'Brien What would compose do?

Conor O'Brien

oh

DJMcMayhem

@ConorO'Brien ? is ternary if, iirc

Conor O'Brien

let's say @ is the composition conjunction. then, f@g (x, y) is the same as f(g(x,y))

HyperNeutrino

01:33

Yes. That makes it a (triadic?) adverb I guess

@ConorO'Brien $.

Conor O'Brien

and that's called...?

HyperNeutrino

<link><link>$ makes a monad out of two links, if not part of an LCC.

It's just called "last two links (if not part of an LCC) as a monad"

Conor O'Brien

it doesn't have a name?

HyperNeutrino

@ConorO'Brien well it's the same as what you described (composition)

Also, ¥ is "last two links (if not part of an LCC) as a dyad".

DJMcMayhem

@ConorO'Brien If F is a monad, then that would just be gf

Conor O'Brien

01:35

I thought dennis had neat little names for each of the commands

HyperNeutrino

Atoms, yes. Quicks, not always.

> unhalve

Another key part of understanding Jelly is vectorization. Does J have that?

Conor O'Brien

that's where the concept game from

HyperNeutrino

oh ok :P

(I know that Jelly was inspired by J and something else IIRC so I was wondering if that came from J)

Conor O'Brien

yup

HyperNeutrino

ah ok

DJMcMayhem

01:37

@HyperNeutrino Correct me if I'm wrong, but if f and g are both dyads, and you call fg dyadically, it'll compute f(g(x, y), x), right?

HyperNeutrino

Yes, but be cautious that that's an unterminated dyad so the behaviour changed based on the next link if present.

DJMcMayhem

o_O

Oliver Ni

tio.run/##MzBNTDJM/Q8ECrq6MQoxNfoK@rq6Cvo1MQA

2

Conor O'Brien

J has much less parsing rules lol

DJMcMayhem

That's weird

Oliver Ni

01:39

oops

@WheatWizard i was born on feb 29....

HyperNeutrino

tio.run/##MzBNTDJM/Q8EAA

hey that's convenient!

Oliver Ni

Starring is messed up

HyperNeutrino

If only this were open still

Conor O'Brien

the cool thing about J is that you can make custom adverbs and conjunctions

   c =: 1 : '(u y) + y'
   *: c 4
20

(*: is "square")

Oliver Ni

the cool thing about C is that you can't make custom classes and stuff

HyperNeutrino

01:44

the cool thing about Jelly is that there aren't even classes anyway so you can't make them :D:D:D

Wait what is =:

Conor O'Brien

@HyperNeutrino "define"

HyperNeutrino

I made a =: in Proton (and a :=) and I was hoping I could have an original idea for once ಠ_ಠ

@ConorO'Brien Ah okay.

Conor O'Brien

@HyperNeutrino bah. there's classes in jelly cuz there's python eval

HyperNeutrino

ok fine :P

Conor O'Brien

and even if not, you can definitely simulate classes

HyperNeutrino

01:49

yeah fine I guess :P I take that back

Qwerp-Derp

02:06

Am I allowed to fanboy about Crystal excessively? Because I might do that :P

Challenger5

What's the shortest way to delete an element from a list Jelly? Like 3,1,4,1,5ç3 -> 3,1,1,5?

HyperNeutrino

@Qwerp-Derp what, here?

And what is Crystal?

Qwerp-Derp

@HyperNeutrino 1. Yes here, and 2. Crystal is a language with Ruby-like syntax, but it compiles to C (therefore it's super fast), and you can write bindings to C super easily

HyperNeutrino

I wonder what the shortest string that pings me is

Qwerp-Derp

And it's got concurrency, Markdown support, it's just nice to program in

HyperNeutrino

02:13

If it's a programming language then it's fine I guess :P

Qwerp-Derp

:P aight

Wheat Wizard

@OliverNi Why did you tell me this?

Qwerp-Derp

I'll make sure to tone it down, but I just really like Crystal (and niche languages in general)

HyperNeutrino

@Challenger5 ĖḢ=¥Ðḟ

dyadic expression

enumerate -> filter out any elements where the first element is equal to the right argument

Alternatively, for fewer unicode chars and more ASCII chars, ĖḢn¥Ðf

enumerate -> filter to keep elements where the first element is not the right argument

Alternatively, ĖḢe¥Ðḟ to filter out elements with a specific index from a list of indices

enumerate -> filter out any elements where the first element is in the right argument (works for single numbers as well)

CMP: Should -a in Proton for a list give the list reversed?

Wheat Wizard

No, thats confusing and weird

HyperNeutrino

02:23

Okay.

Should ~a then? Or should I just make a reverse function?

Wheat Wizard

Probably just make are reverse

or list splicing

I like list splicing

HyperNeutrino

Ah yes, slices. I need those.

DJMcMayhem

Is it a golf language?

HyperNeutrino

Nope.

Qwerp-Derp

@HyperNeutrino GH repo?

HyperNeutrino

02:27

github.com/alexander-liao/proton

If people were to use this language, how desirable would it be to have a[1, 3] get multiple items from a list?

Wheat Wizard

It might be useful

HyperNeutrino

Because it will be a pain to add to the interpreter (the parser already supports it, funnily enough), so if it doesn't seem like it would be used too often, I probably won't add it, at least until later.

Qwerp-Derp

@HyperNeutrino Wouldn't a[1, 3] make more sense to be equal to a[1][3] than a[1], a[3]?

Wheat Wizard

How does your parser work?

HyperNeutrino

@WheatWizard It's similar to ATaco's, and by his description it works much like a greedy regex engine

Wheat Wizard

02:30

@Qwerp-Derp I don't like that, personally

HyperNeutrino

@Qwerp-Derp Well then you could just do a[1][3]

Wheat Wizard

> regex

ew

HyperNeutrino

You're not the first :P

Wheat Wizard

You should use cfgs imo

HyperNeutrino

what is a cfg

Wheat Wizard

02:31

but if the parser works

Context Free Grammar

HyperNeutrino

ah ok

how do those work

Wheat Wizard

They are very simple actually

HyperNeutrino

@WheatWizard it works a lot more nicely than the Positron parser at least :P

Qwerp-Derp

IMO Python has really stupid syntax for class extension...

HyperNeutrino

@WheatWizard Could I have a simple description of how it works?

Wheat Wizard

02:32

You have a bunch of non-terminating symbols (nts) usually represented by capital letters, ...

@HyperNeutrino (typing)

HyperNeutrino

k thx

Wheat Wizard

Each nts maps to a certain number of possibilities that it can be transformed into e.g. S -> $$ | ## .

Which can include other non-terminating symbols as well e.g. S -> (S)

You have a top level symbol, and any string is in the grammar if it can be spanned by some path through the grammar starting from the top level symbol

As an example

The cfg for balanced strings is S -> (S)S | [S]S | {S}S | <S>S | e

where e is the empty string

Does that make sense?

HyperNeutrino

Okay that kind of makes sense.

It sounds suspiciously similar to how my parser works IMO but then I'm probably misunderstanding

Also, I lied about a[2, 3] being hard to implement. In the time it took you to type that explanation, I finished it already, and in the 3 minutes between your last message and my response, I finished making it work for assignment. :P

Wheat Wizard

Being able to parse any cfg is not terribly easy

HyperNeutrino

Oh. I see.

Wheat Wizard

02:40

Its possible your parser works with some subset of cfgs

For example regex can identify some cfgs

HyperNeutrino

Ah okay. That's probably what it is.

Also, how useful do you think a[1:2, 3:4] would be? (comma_expr on slices)

Wheat Wizard

An important thing when writing cfgs is that no string should have two methods of generation

HyperNeutrino

Ah okay. I see.

(Otherwise you get ambiguity? or something?)

Wheat Wizard

@HyperNeutrino That would be pretty useful

@HyperNeutrino That translates to one program having two meanings

which is bad

yes ambiguity is the word

HyperNeutrino

It would be even more of a pain. It would require either spaghetti-coding or an externalized semi-recursive function

Ah okay. Yeah ambiguity is bad.

It's a big problem with most natural languages :P

And with CFGs, ambiguity would require context to discriminate and, well, CFG... :P

Wheat Wizard

02:43

I don't know what your CS background is but if your parser works than it is probably fine, you can rewrite it later (i.e. never)

HyperNeutrino

yay someone agrees with me on the proper definition of later \o/

@WheatWizard My CS background is mostly my dad teaching me some Java, me learning from the internet, and... codegolf... it's partly why my coding is quite bad... I learned a lot from CODEGOLF OF ALL THINGS xD

Also, what's a suitable alternative character for slices? I could try to use : but it would be rather difficult

since it already means different things in other places (foreach)

ppperry

.. maybe

HyperNeutrino

oh yeah probably

Wheat Wizard

@HyperNeutrino Programming != CS

HyperNeutrino

right. well. let me think then.

None

Wheat Wizard

02:45

but there is a good deal of CS on PPCG

ppperry

that's the syntax that MOO uses for ranges and slices

HyperNeutrino

I have very little CS background, if any at all :P

Wheat Wizard

I like ..

HyperNeutrino

@WheatWizard example?

Wheat Wizard

"String"[1,3..5]

HyperNeutrino

02:46

oh ok

So for variable step size, it would be [1..3..5]?

Wheat Wizard

I edited the message for variable step

HyperNeutrino

oh so [1,3..5] is equivalent to [1:3:5] in Python?

Wheat Wizard

That would be [1:5:2]

HyperNeutrino

Also that wouldn't work with Proton unfortunately, because , is already in use so I wouldn't be able to determine which one it means

Conor O'Brien

@Qwerp-Derp Could you link me to a resource about crystal?

HyperNeutrino

02:48

@WheatWizard Oh okay. So [start, second .. last]?

Wheat Wizard

Yeah

HyperNeutrino

cool. that's interesting

officialaimm

I have a knowledge(however limited) about Parse-trees and all, but I thought that needed a lot of work and implemented a parser only by using rercursive functions for my esolang(which is whack btw)..

Wheat Wizard

thats how haskell does ranges

ppperry

that seems just weird to me

HyperNeutrino

02:48

What would be a suitable alternative to that syntax? , .. would not work with Proton for sure

Hey unless I make [a[1], a[3]] equivalent to a[1; 3]

actually that might make more sense with joined slicing

or not

Wheat Wizard

> , .. would not work with Proton for sure

That makes me thing that there is something very wrong with the way you are parsing

HyperNeutrino

so like a[1:5:2] + [a[1], a[3]] -> a[1, 3 .. 5; 1; 3]

Conor O'Brien

oo language design

link to gh repo of language in question?

HyperNeutrino

@WheatWizard No, it's just because , is being used in getitem already..

@ConorO'Brien Proton "Practical" Programming Language, by HyperNeutrino

Wheat Wizard

@HyperNeutrino Could you explain in a little more detail?

ppperry

02:50

This may seem a bit odd, but why not just use plain english

1 to 5 by 3

not much longer

Wheat Wizard

^ I like that too

HyperNeutrino

@WheatWizard a[1, 2] is [a[1], a[2]] so a[1, 2 .. ..... wait actually, that might work?

officialaimm

Guys, what fix-model evaluation do you think is the most fruitful for golfing language? Postfix??

HyperNeutrino

tacit? (not a fix-model but still)

Prefix and Postfix are probably best (and I think the same)

ppperry

How would a[1,2, 3..5] be parsed?

HyperNeutrino

02:52

Right now, it wouldn't.

Wheat Wizard

Tacit is very nice IMO, prefix and postfix are probably the best fixes

ppperry

but cases like that are the danger of reusing symbols.

Qwerp-Derp

@ConorO'Brien Official website

Conor O'Brien

@officialaimm there is no really "best" model; postfix works fine, but so does prefix. I'd say the weakest model is infix, since it almost always necessitates variables, which are (almost always) longer. postfix and prefix are more or less equivalent, but postfix is easier to parse

HyperNeutrino

I like the a to b by c option :D It's very readable (which is the point of Proton) and with numbers it's even decently golfy (1to 2by 3)

@ConorO'Brien and brackets and order of operations

Wheat Wizard

02:52

Haskell can parse [1,2,3..5] as a range

Conor O'Brien

@HyperNeutrino not at all true

HyperNeutrino

no?

Conor O'Brien

no

HyperNeutrino

oh. excuse my ignorance then :P

Conor O'Brien

see pip, I

HyperNeutrino

02:53

@WheatWizard what does that mean

Wheat Wizard

No it can't I lied

HyperNeutrino

rip

I think I'll stick to the a to b by c option.

ppperry

also, may I point out that the OEIS answer-chaining challenge has been auto-protected again

after someone posted and deleted an answer in an already-used language

HyperNeutrino

feature-request: Have Community stop autoprotecting it after it's been unprotected like 100000000000000000000 times

Wheat Wizard

I feel like the one oeis after another has been a bit of a train wreck

ppperry

02:55

how?

HyperNeutrino

@WheatWizard yeah. It got 7 deleted answers in a row :P But it's fun :D

Wheat Wizard

Lost of delted answers and broken chains

HyperNeutrino

especially when you're raging at 12 am because you don't understand the math :D

Anyway I gtg for a bit and I might not be back til tomorrow (wifi cutoff). o/

ppperry

there are actually only 5 deleted answers between 97 and my 98

Conor O'Brien

@HyperNeutrino no offense but this makes me dislike python even more >.>

ppperry

02:56

(although maybe you should count that as 6 since I reused an answer)

Conor O'Brien

@Qwerp-Derp this language looks awesome :o

Transcript for