Lisp Evangelism

why might haskell be accepted?
and why is clojure heretical?

File: IMG_1126.jpg (57KB, 640x480px) Image search: [Google]

57KB, 640x480px

>>56927735
This OP makes it look to an amateur that you know what you're talking about. The fact of the matter is - you're just as lost as I am, friend.

>>56927760
this

op is a serious fag

>>56927760
>Tied to Java
>inb4 ABCL

>>56927761
You have no traveled up the mountain of Enlightenment yet, friendo.

>>56927760
Lisp's close brother Scheme is all about that functional style. Haskell is the pure form of this. Even though it is heretical for disregarding the Holy Paren it still has it's merits in ideology.

File: 1452732608732_0.jpg (624KB, 1242x2208px) Image search: [Google]

624KB, 1242x2208px

>>56927797
But I have. You're leading the way. Have a dank pape for your trouble.

code is data

>>56927797
>Tied to Java
Ah, someone who has never been employed...

>>56927914
In the business world if you wanted to use lisp, then yes, Clojure would be a good choice for compatibility.

Doesn't mean it's a good idea.

>>56927939
>In the business world if you wanted to use lisp, then yes, Clojure would be a good choice for compatibility.

Or, you know, ClojureScript is a thing, and Clojure targets CLR.

Clojure is also a hosted language, in principle it could be very simple to make native. That simply isn't a priority for Clojure customers.

What's your criticism of clojure?

Reminder that Javascript was the most successful implementation of Lisp ever made.

Haskell? They're both functional. So what?
Brainfuck and Python are both imperative.

>>56928061
I would not describe Lisp as being functional. That leaves out a lot of what it means to be Lisp. Lisp is fundamentally a mutiparadigmed language.

Haskell is taking what it means to be functional and pushing it to its limits. Lisp used to be considered functional since historically the idea that functions are data and can be created and referenced arbitrarily was so fundamentally different from the experiences using other languages of the day. I'm not so sure that this is as wild of an idea anymore, so I would refer to Lisp as being mutiparadigmed.

>>56928047
It's way to shitty of a language to come close to the glory of the Holy Paren

>>56928021
Lisp is among other things about doing multi-paradigm programming correctly. Functional is fantastic but Clojure basically is designed to force you into using functional programming only.

It shares ALOT of good things with it's Lisp heritage but misses the point of ultimate freedom done well that Lisp tackles.

>>56928112
> Clojure basically is designed to force you into using functional programming only

You could make that claim because of the way that identity and state is managed. But this is a deliberate choice and a bit of an understatement -- time synchronized concurrency is lock-free and easier to reason about than other approaches to solving these issues (read: Monads) and flexible with regards to data composition. In principle, nothing prevents you from creating traditional objects using Clojure primitives (although this is obviously foolish as more data contention -> more wasted cycles).

Fundamentally, Clojure recognizes the programs take some data and performs some transformation to that data. That data can be anything, even syntax, something Lispers have known for a while. By separating identity and state, very real and practical issues are solved, while simultaneously maximizing freedom.

It would be awesome to see Clojure be taken to the next level by allowing programs to flex across non-temporal environments, i.e. distinct JVMs on distinct machines. In principle, all existing code would be compatible with this due to time-based synchronization.

>>56927760
Because clojure, like mormons, takes existing ideas, filtered through batshit crazy, and run in places nobody sane wants to be (java, or utah)

>>56928612
Kek. Lisp confirmed for the one true programming language.

You have just been accepted into a bishop position at the Church of the Holy Paren

>>56928112
>Functional is fantastic but Clojure basically is designed to force you into using functional programming only.
No more than Scheme or Haskell...

I'm from a Scheme/CL background and don't particularly mind Clojure. It's natural for particular Lisp dialects to be opinionated, that's one of the reasons why the language has lived on. Clojure takes that opinion quite far, but retains the essence of what makes a Lisp.

Just look at things like newLISP or Racket, they're most definitely Lisp dialects with their own biases.

>>56928748
Racket and Clojure are the Protestant equivalents of the Holy Paren Church

And if you're jerking FP so hard then why isn't Common Lisp heretical? It doesn't even require implementations to have essential FP things like tail-call elimination...

>>56928774
Doesn't require but most do. The Church of Holy Paren recognizes the importance of tail call recursion.

>>56928770
Racket is the most Scheme scheme.

>>56928047
It's not homoiconic though.

>>56927735
The core of emacs is coded by C.

>>56929444
I thought it was coded by Stallman.

Who is C?

>>56928061
Brainfuck is memecode for what?

>>56929722
https://en.wikipedia.org/wiki/Brainfuck

>>56929235
I actually agree. But the spirit of scheme still reigns pretty hardcore in racket.

>>56929549
I mean, it is coded "in" C.

Does Racket (PLT Scheme) count?
btw Racket best Scheme

>>56929756
Yes, a very unfortunate truth.
One could certainly rewrite into Lisp compiled to assembly instructions, but there is no need for such a rewrite, and CL is a pragmatic language, not a purist one.

>>56928108
>mutiparadigmed
Shouldn't we call it "omniparadigmed" instead?

File: 1474789172064.jpg (27KB, 500x291px) Image search: [Google]

27KB, 500x291px

>>56929782
C isn't bad language. actually, there's no bad/heretic language. there's only programmers who is so.

>>56929767
It's basically SUPER scheme. So the Holy Church of Paren recognizes it as not scheme but best scheme. (Because scheme is supposed to be completely minimal and Racket violates that but for good reason)

>>56929802
C is a dangerous language. There is not very much that makes C "good". It is simple, and that is just about it. Nearly half of the semantics are "unspecified", and it has petty support for high level abstractions that could be provided at zero cost. C was made to be easy to implement and easy to propagate. C is a virus.

I can't believe you guys aren't all using perl. What kind of /g is this?

i like how with the advent of C++11, C++ has turned into an ugly Lisp

>>56929860
Perl is literally the grossest language in the world. I'd rather program in fucking ASM. Besides all the perl regex specialties are usually ported to other languages almost immedietely.

Perl can die already.

>>56929877
Basically nothing handles PCRE, perl is probably the strongest string handling language that has ever been written.

It's decent at a lot of OTHER things, the CPAN is mostly out of date, there's a lot of awful shit that should not exist.

lisp is slightly overrated

>>56929874
Yes, but minus the symbolic manipulation and code-as-data :(

(templates don't count because you don't have the runtime at your disposal)

>>56930011
that's true. you can do some stuff with templates, but you don't have the entire C++ language. desu the only thing i've seen done with template programming is make a factorial function, and even that simple task looked needlessly complicated.

>>56930045
Right. Technically, templates __are__ Turing-complete, so you can do some insane shit with them if you feel like it. But it will never be as elegant or simple as just giving you the language runtime. When that happens, you have basically achieved lisp.

Modern type systems (like Shen's) or contract systems (like Racket's contracts or Clojure's spec) are so appealing because they offer the full language runtime. Yeah, you have some untypeable expressions, but some languages, like Haskell with a couple extensions or C++, have those anyway. If you're gonna have them, you may as well have the full language runtime at your disposal.

Though I have been a professional C++ developer. Some code is so horrifying it makes you weep, but you can also write some damn good C++ if you are judicious.

>>56929925
It's only overrated in that there is no lisp that is:
-a lisp-1
-with good library support
-without a gc
-without memory management either
-without tons of bugs
One and two almost exist (see: chicken, guile). Usually, when point 1 and 2 are satisfied, point 5 is enough to disqualify it, though I think there's only one lisp at all without a gc (and it uses manual memory management).

>>56929901
What? Virtually everything handles PCRE, even obscure shit languages nobody's ever heard of.

>>56930171
tbf part of the point of Lisp was to abstract you away from those details. Modern advances in GC (like Go's) are leaving less and less room for critique.

>>56930203
Almost everyone handles SOME SUBSET of perl regex, but nothing is anywhere near perl regex. Maybe ruby, I haven't looked hard at their regex libs.

Not to say that most cases need even that much or that you can achieve useful performance with perl tier regex, but as a single string swiss-army-knife you want perl regex hands down.

>>56927777
Quads never lie

>>56930274
>2016
>Not being a fag

I'm OP and at least I'm *current-year*

>>56930270
Wrong.

>>56930246
Go's GC does not employ any new nor modern technology.
Systems like contracts and uniqueness typing can completely eliminate the issue for free, while simultaneously reducing the implementation complexity of the language radically, allowing programs to run without a language runtime, and preventing 90%+ of meaningful bugs from being legal. They're the correct way to go.

>>56927735
LISP is MIT shit

here is some other MIT shit for your enjoyment
http://moralmachine.mit.edu/

>>56930504
>too stupid to get into MIT

thanks for joining us pajeet

>>56930532
stanford is where you go to do CS research

MIT is where you go to do trolley problems

>>56930045
> desu the only thing i've seen done with template programming is make a factorial function
look at eigen for a serious use of templates

if you have numerical vectors A = B + C + D, then in naive C++ you'll instantiate a temporary vector (B+C) to add to D. eigen's metaprogramming instead instantiates a constructor A(B,C,D) that does the correct B[x] + C[x] + D[x] thing internally

>>56930570
CSAIL > everything at stanford

enjoy being the next brogrammer at snapchat