The Nineteenth Byte

ASCII-only

12:00 AM

@Pavel well... the brain is a neural network

computers use an artificial neural network

Pavel

yes

that was indeed the point

Veskah

Also, if I vanish mysteriously, it means I was too close to the truth and Dennis dispatched TIObots to silence me.

Downgoat

12:59 AM

for the truth machine catalog do answers need to be full programs or are functions ok

no idea how IO should work in VSL yet...

Pavel

@Downgoat What do you mean?

Downgoat

1:22 AM

@Pavel VSL can't take input from STDIN for the truth machine challenge yet

I suppose it technically could but not really

Pavel

When you're asking how it should work do you mean what the functions for it should look like or what

lirtosiast

@LeakyNun Is this even computable?

Pavel

@Downgoat please make input reading support line editing

BMO

@lirtosiast You could use it to solve the Halting problem, but "in a language of a choice" means you can pick one where the Halting problem can be solved ;P

Downgoat

@Pavel example?

ASCII-only

1:29 AM

@BMO so... one of ais's

Downgoat

@Pavel both as in things like buffering, async behavior, etc

BMO

@ASCII-only Do they only have non-TC langs?

ASCII-only

@BMO no, but they have plenty of uncomputable ones

uncomputable as in too powerful to compute :P

lirtosiast

@BMO can't we just choose the trivial language then?

BMO

@ASCII-only No you'd need easier languages which are not even TC

Pavel

1:33 AM

@Downgoat readline

BMO

@lirtosiast I guess, but they still need to be able to do computations on natural numbers, so it's not soo easy

Pavel

The c library

ASCII-only

@BMO why not more powerful

Pavel

@Downgoat readln with optional filehandle argument

lirtosiast

@BMO Unless I'm missing something it seems like the challenge is ill-defined

BMO

1:33 AM

Because you can't even solve it for TC languages, so you'll want a less powerful model where you can solve the Halting problem

Pavel

Just do it synchronously by default

lirtosiast

what is the class of "computations on natural numbers" defining the minimum power of such a language?

ASCII-only

@BMO no, IIRC ais has languages powerful enough to solve the halting problem. the only problem is that they're too powerful to create an interpreter for

BMO

@ASCII-only Aah, you're thinking of a language to solve the problem in :) I was thinking of what language's code complexity to determine

ASCII-only

hmm. well i'm assuming the language can (theoretically) be both

lirtosiast

1:38 AM

The language you solve it in has to be more powerful than the language you solve it on

BMO

@lirtosiast Yes, unless you define some weird language which cannot exist

Is there a computation model between FSM and push-down automata?

Downgoat

@Pavel I thought about this but linking GNU readline is non-trivial dependency

Pavel

@Downgoat Grab the source and shove it into vsl?

Then you don't need to link

lirtosiast

2:13 AM

Has the idea of duels ever come to anything, and does the community still think it a good idea?

DJMcMayhem

@lirtosiast There have been several duels before. I think most of the community thinks they're a good idea, there's just nothing official about them

lirtosiast

So there's no official list anywhere?

Also, any duels between golfing languages?

ASCII-only

2:37 AM

~~Charcoal and Canvas would be a good one~~

Veskah

A duel between Jelly and a .45

DJMcMayhem

2:58 AM

I've gotten in a lot of unofficial V/osabie duels, but I'm pretty sure osabie has the upper hand %98+ of the time

ASCII-only

yeah. O5AB1E beat Charcoal in ASCII-art once :(

lirtosiast

Charcoal spends a lot of its codepage on non-ASCII art things now, no?

ASCII-only

kinda but not really

there are a lot more operators now, i guess

but you kinda need some processing capabilities in addition to drawing commands anyway

Downgoat

@Pavel readline is pretty huge with history and all

I can probably just enable cook mode

Leaky Nun

3:22 AM

@lirtosiast see, the point of my cmc was to create a discussion, which it succeeded to :p

BMO

@LeakyNun But the real question is, did you have a special model in mind where you would have considered a solution to be interesting?

Also does nobody know of sth. between a FSM and push-down automata?

Veskah

The real answer is the friends we made along the way

Pavel

4:06 AM

@Downgoat history is good tho

lirtosiast

@BMO I barely even know what those are, but my intuition for the power of grammars / models of automata is that there are always more in between

BMO

@lirtosiast Maybe there is, I don't know

My intuition tells me the opposite :S

Thinking that you can do pretty weird things (like allowing non-determinism) and ending up with the same model or the Church-Turing thesis

lirtosiast

Hmm

Allowing nondeterminism seems like an obvious thing though

BMO

4:21 AM

If you're familiar with the concepts..

I don't think it was obvious to me that an NFA is equivalent to an DFM before knowing the proof

lirtosiast

Maybe some class B can be generated by saying "C but you can't do x" or "A but you can also do x" where x is some particular operation that isn't C-complete

BMO

I don't think such an x always exists?

lirtosiast

Nested word

In computer science, more specifically in automata and formal language theory, nested words are a concept proposed by Alur and Madhusudan as a joint generalization of words, as traditionally used for modelling linearly ordered structures, and of ordered unranked trees, as traditionally used for modelling hierarchical structures. Finite-state acceptors for nested words, so-called nested word automata, then give a more expressive generalization of finite automata on words. The linear encodings of languages accepted by finite nested word automata gives the class of visibly pushdown languages. The...

I think I found on on Wikipedia?

BMO

@lirtosiast Ah nice, seems like that's the case

BMO

4:40 AM

That's a nice find :)

New Main Posts

5:07 AM

0

Q: Modular distance

Dropping this question here because I've been stuck for hours now. I want to correctly calculate the maximum modular distance between numbers in a list. Modular distance is the smallest distance two numbers would be apart from each other on a clock with x markers. Take a regular clock (twelve mar...

math

wizzwizz4

12:36 PM

Does anyone know if you can change the name / email address associated with a Heroku App account?

primo

4:04 PM

(haskell) for the default guard, is there anthing shorter than f n|0<1=... ?

H.PWiz

Not in general

Sometimes you can skip it with a pattern match below

primo

f n|n<9=0;f n=...
f n|n<9=0|0<1=...

seems shorter just to define the default separately

or wait, no difference :/

H.PWiz

There is also that you can put some computation in the guard: f n|a<-n+2=...

primo

that is interesting

H.PWiz

I'm waiting for your λ answer in Haskell now ;). I've nearly caught up to you in julia

primo

4:14 PM

i'll let you know when i pass you :p

lirtosiast

6:54 PM

Does any kind of spreadsheet of golfing languages exist listing their various advantages and disadvantages?

Adám

@lirtosiast Yes. It was announced here in chat, I think. Let me see if I can find it.

@lirtosiast github.com/ETHproductions/golfing-langs

lirtosiast

ty

I wonder if we could add columns to that like "number of builtins (tens, hundreds, thousands", "syntactical efficiency (low-high)", "variance in code length (low-high)"

oh @Adám I also have a question for you

DJMcMayhem

That page looks absolutely horrific on mobile

lirtosiast

I remember reading something about the original motivation of trains being that they were somehow theoretically most concise? But I can't find it anywhere

do you know if it's true that the problem of compressing the function tree was studied in detail?

primo

7:35 PM

@H.PWiz your move

H.PWiz

7:45 PM

Nice. Let me think of something clever

Turns out I didn't need anything clever :p

Adám

@lirtosiast No, the original motivation for trains (specifically forks) was as a means of expressing what in TMN (Traditional Mathematical Notation) is denoted f+g. Note that + here acts exactly like an APL operator in that it derives a new function from f and g just like . does in f.g and ∘ does in f∘g. So Iverson played with each (scalar) function having a companion operator (its symbol was the scalar function with an overbar, e.g. + had ∓←{⍺←⊢ ⋄ (⍺ ⍺⍺ ⍵)+(⍺ ⍵⍵ ⍵)}).

lirtosiast

Hmm

Adám

Naturally, this lends itself well to an operator that takes three operands, which APL operators cannot. Eventually, it dawned on him that f g h could be introduced as a functional form.

lirtosiast

And this was when, the 70s, before J?

primo

drat

H.PWiz

7:55 PM

I had a=<<b.c<$> in my program, and golfed it a.b.c=<<

primo

yours is a lot more complex than mine

Adám

@lirtosiast Late '80s, right before "J".

primo

i don't have a single composition >_<

lirtosiast

> Iverson and Whitney have made several earlier suggestions of ways to achieve what the fork form provides: the scalar operators of [Iv78], [Iv79a], [Iv 79b], the til operator of [Iv82], the union and intersection conjunctions of [Iv87], and the yoke adverb of [Iv88]. Benkard [Bk87] has also suggested a way to achieve the meaning of this form, in his proposal for ↑g/(f h)⍺ ⍵ , using the notion of function pair (↑ is APL2’s first function).

H.PWiz

@primo The complex stuff is me trying to compress the polynomial. I only started using things outside ascii this afternoon (GMT)

lirtosiast

7:57 PM

Huh, I assume scalar operators are the overbar ones? What are til, union and intersection conjunctions, and yoke?

H.PWiz

It only saved a dozen or so bytes

primo

same

H.PWiz

Have you seen this? chat.stackexchange.com/transcript/message/48220543#48220543

Adám