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 »

Zombywuf: Mar 29, 2008

So the correct answer is Prolog, cool.

# ? Dec 11, 2012 02:04

Adbot: ADBOT LOVES YOU

# ? May 9, 2024 11:29

trex eaterofcadrs: Jun 17, 2005; My lack of understanding is only exceeded by my lack of concern.

Zombywuf posted:

So the correct answer is Prolog, cool.

clojure + core.logic

# ? Dec 11, 2012 02:20

MeruFM: Jul 27, 2010

bash

# ? Dec 11, 2012 02:34

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

Zombywuf posted:

So the correct answer is Prolog, cool.

if software = data structures + algorithms

algorithms = logic + control :v:

# ? Dec 11, 2012 02:39

Meiwaku: Jan 10, 2011; Fun for the whole family!

trex eaterofcadrs posted:

clojure + core.logic

# ? Dec 11, 2012 02:41

HORATIO HORNBLOWER: Sep 21, 2002; no ambition,
no talent,
no chance

mutability sucks and is stupid immutable structures go 4ever

# ? Dec 11, 2012 02:50

Opinion Haver: Apr 9, 2007

MononcQc posted:

immutable data structures are pretty nice to have but also getting well-delimited mutability is a nice tool. I do prefer a clear delimitation between both modes than only mutable if available though.

IORef/State :colbert:

0xB16B00B5 posted:

look jsut loving decide what X is already stop changing your mind later int he program

i mean whatever it is we know it gon give it to ya

# ? Dec 11, 2012 02:52

MononcQc: May 29, 2007

Monad stuff, independent data structure allowing modifications or not (what Schemes tend to do now, specifically Racket), clearly-defined semantics about mutability like what Clojure does, strict use of const stuff in Cs, uniqueness types, whatever.

I think (and have no data to back it up) that it's fairly important to make mutation very explicit when it happens, to have well-defined semantics between "this is safe" and "this isn't". If you can differentiate between state and identity the way Clojure does it then it's even better. I don't particularly care how it's done, as long as you can do it reliably.

It's nice to be able to have a reference to some data structure and know you can make idempotent operations on it, or that it's impossible and you should just not assume you can.

I'd generally go for immutable stuff by default and then get mutation or destructive updates when I actually have a real need to, either through performance requirements, or semantic ones. Then again, I work in systems and languages that give this to me by default, so that became my defacto thinking mode.

# ? Dec 11, 2012 03:01

Sang-: Nov 2, 2007

Tiny Bug Child posted:

it is

where would the ability to perform this on types be useful

union types are really useful

code:

object Bar {
   def foo(xs: (String or Int)*) = xs foreach {
      case _: String => println("str")
      case _: Int => println("int")
   }
}

(note you can't do this in Scala)

# ? Dec 11, 2012 03:38

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

Tiny Bug Child posted:

where would the ability to perform this on types be useful

ugh, academics, things are all the same type in a real programming language

# ? Dec 11, 2012 03:47

Tiny Bug Child: Sep 11, 2004; Avoid Symmetry, Allow Complexity, Introduce Terror

in an ideal programming language, sure, but like an ideal gas no such thing truly exists

# ? Dec 11, 2012 04:02

wins32767: Mar 16, 2007

Otto Skorzeny posted:

now you have 32767 problems

Nah, we cool

# ? Dec 11, 2012 04:03

Tiny Bug Child: Sep 11, 2004; Avoid Symmetry, Allow Complexity, Introduce Terror

tef posted:

ugh, academics, things are all the same type in a real programming language

as a way of zeroing in on the perfect platonic ideal of a programming language, i've been kicking around this idea to help further remove the meaningless distinctions between "types": make everything an array. primitives as we currently know them are just arrays of length 1

# ? Dec 11, 2012 04:08

MononcQc: May 29, 2007

should go back to bits and bytes instead. More general.

# ? Dec 11, 2012 04:08

Tiny Bug Child: Sep 11, 2004; Avoid Symmetry, Allow Complexity, Introduce Terror

like if you try to do this:

$it = "string";
$it[] = "string2";

php bitches at you, when it should turn $it into array("string", "string2")

# ? Dec 11, 2012 04:09

jooky: Jan 15, 2003

Tiny Bug Child posted:

like if you try to do this:

$it = "string";
$it[] = "string2";

php bitches at you, when it should turn $it into array("string", "string2")

*nods*

# ? Dec 11, 2012 04:10

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

van der waals of abstraction

# ? Dec 11, 2012 04:55

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

Tiny Bug Child posted:

as a way of zeroing in on the perfect platonic ideal of a programming language, i've been kicking around this idea to help further remove the meaningless distinctions between "types": make everything an array. primitives as we currently know them are just arrays of length 1

what about if you just made everything a function. so like primitives would just be a function that always returns the same value

:2bong:

# ? Dec 11, 2012 05:10

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

Tiny Bug Child posted:

like if you try to do this:

$it = "string";
$it[] = "string2";

php bitches at you, when it should turn $it into array("string", "string2")

shouldn't it turn it into "stringstring2" :confused:

# ? Dec 11, 2012 05:11

double sulk: Jul 2, 2010

tef posted:

shouldn't it turn it into "stringstring2"

php

# ? Dec 11, 2012 05:15

Tiny Bug Child: Sep 11, 2004; Avoid Symmetry, Allow Complexity, Introduce Terror

tef posted:

shouldn't it turn it into "stringstring2"

nah that would be the . operator

# ? Dec 11, 2012 05:17

rotor: Jun 11, 2001; classic case of pineapple derangement syndrome

JawnV6 posted:

van der waals of abstraction

this joke seems kind of forced.

# ? Dec 11, 2012 08:12

milquetoast child: Jun 27, 2003; literally

I sort of want to learn python.

My "code" skills are ~1yr of SQL dba, and a few years of HTML. I took AP comp sci in high school 12 years ago in C++ and got a 3 on the exam, but I was written and I remember nothing.

Where do I start?

# ? Dec 11, 2012 08:49

rotor: Jun 11, 2001; classic case of pineapple derangement syndrome

JosephStalinVEVO posted:

I sort of want to learn python.

Where do I start?

how about here:

http://www.learnpython.org/

# ? Dec 11, 2012 08:57

homercles: Feb 14, 2010

Sang- posted:

union types are really useful
code:
object Bar {
   def foo(xs: (String or Int)*) = xs foreach {
      case _: String => println("str")
      case _: Int => println("int")
   }
}
(note you can't do this in Scala)

I thought this would be easily doable via implicit functions doing the mapping but the answers here do it through even nicer means.

http://stackoverflow.com/questions/3508077/does-scala-have-type-disjunction-union-types
http://www.chuusai.com/2011/06/09/scala-union-types-curry-howard/

this does a thing, somehow. really batshit crazy stuff i wish i understood

# ? Dec 11, 2012 08:59

milquetoast child: Jun 27, 2003; literally

rotor posted:

how about here:

http://www.learnpython.org/

Ok but should I buy a new retina mbp so I can code right!?

# ? Dec 11, 2012 09:02

rotor: Jun 11, 2001; classic case of pineapple derangement syndrome

JosephStalinVEVO posted:

Ok but should I buy a new retina mbp so I can code right!?

no, a thinkpad.

# ? Dec 11, 2012 09:05

homercles: Feb 14, 2010

really like wtf i'm going to read more scala and come back to that crazy poo poo

# ? Dec 11, 2012 09:06

rotor: Jun 11, 2001; classic case of pineapple derangement syndrome

homercles posted:

I thought this would be easily doable via implicit functions doing the mapping but the answers here do it through even nicer means.

http://stackoverflow.com/questions/3508077/does-scala-have-type-disjunction-union-types
http://www.chuusai.com/2011/06/09/scala-union-types-curry-howard/

this does a thing, somehow. really batshit crazy stuff i wish i understood

i dont think strong typing is worth the effort

# ? Dec 11, 2012 09:10

homercles: Feb 14, 2010

maybe in 2.10 you could implement such a feature more explicitely with macros

it would still be butt ugly

# ? Dec 11, 2012 09:25

Workaday Wizard: Oct 23, 2009; by Pragmatica

quote:

def size(x : Int ∨ String) = x match { case i : Int => i case s : String => s.length }

zoom

quote:

def size(x : Int ∨ String) = x match {

enhance

quote:

∨

:∨

# ? Dec 11, 2012 09:28

Workaday Wizard: Oct 23, 2009; by Pragmatica

D34TH ∀

# ? Dec 11, 2012 09:30

qntm: Jun 17, 2009

tef posted:

what about if you just made everything a function. so like primitives would just be a function that always returns the same function

# ? Dec 11, 2012 12:47

jony neuemonic: Nov 13, 2009

JosephStalinVEVO posted:

I sort of want to learn python.

My "code" skills are ~1yr of SQL dba, and a few years of HTML. I took AP comp sci in high school 12 years ago in C++ and got a 3 on the exam, but I was written and I remember nothing.

Where do I start?

there's an mit course on itunes u that uses python. that seemed fine but it might be too intro-level for you.

# ? Dec 11, 2012 13:34

double sulk: Jul 2, 2010

learn ruby instead

# ? Dec 11, 2012 14:01

0xB16B00B5: Aug 24, 2006; by Y Kant Ozma Post

web "developer"status: sass owns so much porting everything over to it

# ? Dec 11, 2012 14:05

prefect: Sep 11, 2001; No one, Woodhouse.
No one.; Dead Man’s Band

tef posted:

what about if you just made everything a function. so like primitives would just be a function that always returns the same value

Isn't that what Lisp is like? Or OCaml or one of those functional things?

VVVV :saddowns:

prefect fucked around with this message at 14:33 on Dec 11, 2012

# ? Dec 11, 2012 14:12

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

# ? Dec 11, 2012 14:30

MononcQc: May 29, 2007

unwarranted effort post:

Lisp and other similar languages are based on the idea of lambda calculus (the function-based everything), but for practical reasons, ditch a lot of "everything is a function" for efficiency reasons.

For example, by default, Scheme'S lists are the usual 'cons cell' data structure. You can think of it as a C struct with two pointers (current and next), which are pointers to the current value or the following one:

code:

> (cons 'a 'b)
'(a . b)
> (cons 'a '())
'(a)
> (cons 'a (cons 'b '()))
'(a b)
> (list 'a 'b)
'(a b)
> (car (cons 'a (cons 'b '())))
'a
> (cdr (cons 'a (cons 'b '())))
'(b)

A pointer to '() is similar to a pointer to NULL, pointing towards the end of the list. Now that's the practical implementation, but nothing keeps you from redefining lists as functions:

code:

> (define (cons x y) (lambda (f) (f x y)))
> (define (car l) (l (lambda (x y) x)))
> (define (cdr l) (l (lambda (x y) y)))
> (cons 'a 'b)
#<procedure>
> (cons 'a '())
#<procedure>
> (cons 'a (cons 'b '()))
#<procedure>
> (car (cons 'a (cons 'b '())))
'a
> (cdr (cons 'a (cons 'b '())))
#<procedure>
> (car (cdr (cons 'a (cons 'b '()))))
'b

So I'm able to replace language primitives with actual functions and reimplement list operations. Now what you've got to do is keep doing that everywhere, which takes long and will give you a very slow language.

Lambda calculus (untyped) works that way from the beginning, depending on if you allow the notation for anonymous functions to take more than one argument. If not, you need to work your way through currying to get something equivalent.

Languages in the ML family (ML, SML, OCaml, Haskell, etc.) instead base themselves on typed lambda calculus, which is significantly more complex, but allows the presence of type systems similar to what these languages have (they tend to pick Hindley-Milner).

# ? Dec 11, 2012 14:36

Adbot: ADBOT LOVES YOU

# ? May 9, 2024 11:29

MononcQc: May 29, 2007

reimplementing lists as functions to talk about Schemes is my most common use of Schemes

# ? Dec 11, 2012 14:41

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

«‹›1784 »