p-lang thread: (now (have you (problems two)))

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > p-lang thread: (now (have you (problems two)))

«‹›1784 »

Cybernetic Vermin: Apr 18, 2005

considering the vast amount of time people waste on new languages one would think someone would get around to a fast modern (e.g. integrate with gpu compute, maybe just parasite on numpy) apl/k/q take.

not that it'd be terribly useful, but i'd enjoy it a lot more than another weird midpoint between rust and c

Cybernetic Vermin fucked around with this message at 15:12 on Aug 29, 2019

# ? Aug 29, 2019 15:06

Adbot: ADBOT LOVES YOU

# ? Jun 6, 2024 09:40

animist: Aug 28, 2018

it's called pytorch

# ? Aug 29, 2019 15:42

Cybernetic Vermin: Apr 18, 2005

animist posted:

it's called pytorch

that's where i'm at now, but it is very much some array stuff added into an otherwise lovely language.

e: though tbqh i guess a small amount of syntactic sugar added would go a long way, the array underpinnings and overall ecosystem is after all really great.

Cybernetic Vermin fucked around with this message at 16:21 on Aug 29, 2019

# ? Aug 29, 2019 16:06

fritz: Jul 26, 2003

Cybernetic Vermin posted:

considering the vast amount of time people waste on new languages one would think someone would get around to a fast modern (e.g. integrate with gpu compute, maybe just parasite on numpy) apl/k/q take.

not that it'd be terribly useful, but i'd enjoy it a lot more than another weird midpoint between rust and c

theres this thing : https://github.com/Co-dfns/Co-dfns

# ? Aug 29, 2019 16:52

animist: Aug 28, 2018

Cybernetic Vermin posted:

that's where i'm at now, but it is very much some array stuff added into an otherwise lovely language.

e: though tbqh i guess a small amount of syntactic sugar added would go a long way, the array underpinnings and overall ecosystem is after all really great.

I mean there's TVM / tiramisu (http://tiramisu-compiler.org/) which are technically languages in the way LLVM IR is a language... but those are embedded in Python / C++ respectively

bad days for inscrutable array languages tbh

# ? Aug 29, 2019 17:17

Athas: Aug 6, 2007; fuck that joker

animist posted:

bad days for inscrutable array languages tbh

hello

# ? Aug 29, 2019 17:40

Carthag Tuek: Oct 15, 2005; Tider skal komme,
tider skal henrulle,
sl�gt skal f�lge sl�gters gang

Athas posted:

hello

not written in runes, banned until ragnarok

# ? Aug 29, 2019 19:54

champagne posting: Apr 5, 2006; YOU ARE A BRAIN
IN A BUNKER

Cybernetic Vermin posted:

considering the vast amount of time people waste on new languages one would think someone would get around to a fast modern (e.g. integrate with gpu compute, maybe just parasite on numpy) apl/k/q take.

not that it'd be terribly useful, but i'd enjoy it a lot more than another weird midpoint between rust and c

don't worry you can have coworkers waste time with old niche languages as well

# ? Aug 29, 2019 21:51

eschaton: Mar 7, 2007; Don't you just hate when you wind up in a store with people who are in a socioeconomic class that is pretty obviously about two levels lower than your own?

Cybernetic Vermin posted:

considering the vast amount of time people waste on new languages one would think someone would get around to a fast modern (e.g. integrate with gpu compute, maybe just parasite on numpy) apl/k/q take.

not that it'd be terribly useful, but i'd enjoy it a lot more than another weird midpoint between rust and c

I never understood why there wasn�t just an S-expression conversion for APL

like, you get both straightforward readability and the ability to compile the full program however will be most efficient for your circumstances, what�s not to like?

# ? Aug 30, 2019 10:28

Athas: Aug 6, 2007; fuck that joker

eschaton posted:

I never understood why there wasn�t just an S-expression conversion for APL

like, you get both straightforward readability and the ability to compile the full program however will be most efficient for your circumstances, what�s not to like?

Parsing APL is type-dependent. Specifically, parsing e.g. 'x y z' depends on whether 'y' has been defined as a unary or binary function (called "monadic" and "dyadic" in APL), as an operator (higher-order function), or maybe not as a function at all. Fortunately, this is only really a problem with dfns (sort of like anonymous functions), and these can take at most only four parameters (always named α, ω, α, ω) and so you can just generate all of them and dispatch once you know the types at runtime.
APL is entirely about the notation. The overloaded semantics are not nearly as important, and often inhibit efficient compilation.
APL is not actually particularly easy to compile, and real-world APL tends to be quite imperative.

APL is one of those languages that is best admired from afar. It has tons of warts once you get close. I have hacked on a parallelizing APL compiler and I don't really want to go back to it.

# ? Aug 30, 2019 13:28

ColTim: Oct 29, 2011

k/q has what I assume to be a pretty-lispy parse tree notation which ends up getting used in functional selects/deletes/updates, e.g. "x+y" is equivalent to "(+;x;y)", the latter of which is a generic list but which can also be evaluated/run using "value" or monadic "."

# ? Aug 30, 2019 14:54

Progressive JPEG: Feb 19, 2003

can someone release a fortran package management system named forball? tia

# ? Aug 30, 2019 21:41

Plank Walker: Aug 11, 2005

Progressive JPEG posted:

can someone release a fortran package management system named forball? tia

Interestingly, no. At this level, the market demands non-linear value. We go down to a single ball.

# ? Aug 30, 2019 22:53

eschaton: Mar 7, 2007; Don't you just hate when you wind up in a store with people who are in a socioeconomic class that is pretty obviously about two levels lower than your own?

one cross-plang package manager to rule them all

call it balls

# ? Aug 31, 2019 01:04

distortion park: Apr 25, 2011

ColTim posted:

k/q has what I assume to be a pretty-lispy parse tree notation which ends up getting used in functional selects/deletes/updates, e.g. "x+y" is equivalent to "(+;x;y)", the latter of which is a generic list but which can also be evaluated/run using "value" or monadic "."

I'm not sure that making a fancy k/q with gpu acceleration and stuff would work. Half of the good stuff about it comes from its simplicity

# ? Aug 31, 2019 09:05

Cybernetic Vermin: Apr 18, 2005

pointsofdata posted:

I'm not sure that making a fancy k/q with gpu acceleration and stuff would work. Half of the good stuff about it comes from its simplicity

arrays being marked with transient but explicit properties which optimize some operations is sort of a core mechanism of the language though, so having a `gpu#x list would fit in pretty well.

e: to clarify to anyone curious, in k/q you go e.g. `s#x to give the list 'x' the 'sorted' attribute (the sorting primitives produce lists with this attribute), making e.g. lookups in the list become log(n) operations. the attribute is preserved by most operations which have no effect on order, but is silently lost by the rest. there's a few such attributes (two other kinds of indices, simple rle compression, etc.), and `gpu# would fit in pretty well.

another thing about array languages though is that it is a very commercial area, in that the decent apl implementations (and this includes e.g. the above co-dfns above being hosted on dyalog), k and q are all extremely expensive licenses, and j is gpl (with the option, again, of commercial licenses) landing you in the usual quagmire when you want to make something a library (rather than everything being a monolithic program).

Cybernetic Vermin fucked around with this message at 09:42 on Aug 31, 2019

# ? Aug 31, 2019 09:25

JawnV6: Jul 4, 2004; So hot ...

https://twitter.com/hynek/status/1168252462513475589

# ? Sep 2, 2019 16:05

champagne posting: Apr 5, 2006; YOU ARE A BRAIN
IN A BUNKER

JawnV6 posted:

https://twitter.com/hynek/status/1168252462513475589

it�s not wrong but there�s a couple of steps between terrible programming and genocide enabling

# ? Sep 2, 2019 16:15

Xarn: Jun 26, 2015

This sort of nuclear takes is why I visit this thread. Do you think he belongs to the "JS is perfectly fine language for the kernel" people, or to the "obviously people should've been writing Rust 40 years ago, also what is unsafe" people?

# ? Sep 2, 2019 16:31

Nomnom Cookie: Aug 30, 2009

rust was appropriate 40 years ago but now we have modern tools like js

# ? Sep 2, 2019 16:37

akadajet: Sep 14, 2003

JawnV6 posted:

https://twitter.com/hynek/status/1168252462513475589

pictured: a genocide enabling crab
https://www.youtube.com/watch?v=Snr113r5ocY

# ? Sep 2, 2019 16:56

animist: Aug 28, 2018

i only use memory-safe languages, like COBOL

# ? Sep 2, 2019 17:54

Vomik: Jul 29, 2003; This post is dedicated to the brave Mujahideen fighters of Afghanistan

why does everyone in Silicon Valley think they�re so important?

# ? Sep 2, 2019 18:20

Sapozhnik: Jan 2, 2005; Nap Ghost

more money = better than

# ? Sep 2, 2019 18:32

pokeyman: Nov 26, 2006; That elephant ate my entire platoon.

because they�re temporarily getting paid a lot of money, and that�s how they determine importance?

# ? Sep 2, 2019 18:32

Subjunctive: Sep 12, 2006; ✨sparkle and shine✨

because Silicon Valley gets a lot of attention for virtually everything

# ? Sep 2, 2019 18:32

mystes: May 31, 2006

Vomik posted:

why does everyone in Silicon Valley think they�re so important?

Because that's how capitalism works.

# ? Sep 2, 2019 21:20

animist: Aug 28, 2018

coq is cool but I have no goddamn idea how it works

# ? Sep 12, 2019 23:27

Captain Foo: May 11, 2004; we vibin'
we slidin'
we breathin'
we dyin'

animist posted:

coq is cool but I have no goddamn idea how it works

you just jack it around until you get some output

# ? Sep 12, 2019 23:34

Lutha Mahtin: Oct 10, 2010; Your brokebrain sin is absolved...go and shitpost no more!

it takes a bit of skill to get the desired effect tho

# ? Sep 13, 2019 00:59

rjmccall: Sep 7, 2007; no worries friend; Fun Shoe

the lambda cube is like a chimichanga

# ? Sep 13, 2019 02:56

Zlodo: Nov 25, 2006

Captain Foo posted:

you just jack it around until you get some output

# ? Sep 13, 2019 11:06

Tanners: Dec 13, 2011; woof

Captain Foo posted:

you just jack it around until you get some output

# ? Sep 13, 2019 13:58

tef: May 30, 2004; -> some l-system crap ->

Boiled Water posted:

it�s not wrong but there�s a couple of steps between terrible programming and genocide enabling

yeah the facebook interview drags out

# ? Sep 15, 2019 14:29

tef: May 30, 2004; -> some l-system crap ->

C bugs in lovely routers, etc, sure enough but saying that somehow Rust gives you more safety against a hostile nation state is a fundamental misunderstanding of the threat model

# ? Sep 15, 2019 14:31

Subjunctive: Sep 12, 2006; ✨sparkle and shine✨

tef posted:

C bugs in lovely routers, etc, sure enough but saying that somehow Rust gives you more safety against a hostile nation state is a fundamental misunderstanding of the threat model

is this �you�ll never prevent all the bugs so why try to prevent any�?

# ? Sep 15, 2019 16:16

Suspicious Dish: Sep 24, 2011; 2020 is the year of linux on the desktop, bro; Fun Shoe

9/11 would have been prevented if the plane had proper memory segmentation

# ? Sep 15, 2019 16:17

Suspicious Dish: Sep 24, 2011; 2020 is the year of linux on the desktop, bro; Fun Shoe

the plane should have had total separation of code and data, like my favorite programming language which exemplifies total separation of code and data, lisp

# ? Sep 15, 2019 16:19

Workaday Wizard: Oct 23, 2009; by Pragmatica

if you want to use that analogy then you should compare it to having the pilots isolated from the common filth

# ? Sep 15, 2019 17:19

Adbot: ADBOT LOVES YOU

# ? Jun 6, 2024 09:40

Jerry Bindle: May 16, 2003

Captain Foo posted:

you just jack it around until you get some output

# ? Sep 15, 2019 17:28

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > p-lang thread: (now (have you (problems two)))

«‹›1784 »