>>121 You should define pi as the way to compute it
Symbolic computation everywhere is unnedeed and impractical. Lousy approximations are unneeded, impractical and harmful. They corrupt the art/engineering/magic of Computer Science into arcane superstition. See http://www.loper-os.org/?p=41
Name:
Anonymous2013-06-24 12:53
>>122 Why would anybody want to do that?
To prove that Haskell can at least act as a substitution.
Name:
Anonymous2013-06-24 13:04
>>124
Haskell isn't meant to be the language for everything on Earth, particularly not for low-level, high-performance infrastructural stuff.
Name:
Anonymous2013-06-24 13:10
>>125 Haskell isn't meant to be the language for everything on Earth,
That is why Haskell is inherently worse than Lisp or even C++.
Name:
Anonymous2013-06-24 13:19
>>126
You're wrong because
a) Lisp isn't a single language, it's a LOT of languages
b) C++ isn't good for everything, it isn't used for everything, it loses out to other languages in a lot of areas
Name:
Anonymous2013-06-24 13:26
>>127 Lisp isn't a single language, it's a LOT of languages
Lisp is a single language, which can be adapted to any domain, including '(bare words english language).
C++ isn't good for everything, it isn't used for everything, it loses out to other languages in a lot of areas
Yet C++ has more practical value than Haskell. At least C++ can be self hosting.
Name:
Anonymous2013-06-24 13:29
>>128 Lisp is a single language, which can be adapted to any domain
Bullshit. (defmacro) or (define)? (defun) or (fn)? Is there CLOS or not? How about Racket, where you can switch languages, choosing any one of different Rackets? Can you do sequent calculus in Scheme like you can in Shen? Etc etc.
What's the advantage of being self-hosting? Haskell is a more high-level language and doesn't have to be self-hosting to be useful.
Name:
Anonymous2013-06-24 14:20
linux is superior to windows and mac os x that's for sure
Name:
Anonymous2013-06-24 14:38
>>129 Bullshit. (defmacro) or (define)? (defun) or (fn)?
different names for same concept.
Is there CLOS or not?
CLOS is a library.
How about Racket, where you can switch languages, choosing any one of different Rackets?
That is a bad feature. Racket is a little bloated.
What's the advantage of being self-hosting?
It proves that language is general enough and doesn't need any external tool.
Haskell is a more high-level language and doesn't have to be self-hosting to be useful.
There is nothing Haskell can do to be useful. Although you can write a postmortem obituary on Haskell.
It continues to amaze me how bad Haskell is at processing strings. One of the reasons I wanted to learn Haskell was to be able to write short, dynamic-language-like programs that execute fast once compiled. Somehow rather, Haskell has failed to deliver on its promise of bare metal speed. (http://honza.ca/2012/10/haskell-strings)
Name:
Anonymous2013-06-24 15:05
>>131
But they're still different fucking languages. Just like Java is different from C# even though they have a lot of common concepts. E.g. Racket's (define-syntax-rule) and (define-syntax) work differently from CL's (defmacro).
No, CLOS is not a library, it's a part of the language specification.
Wait... you're Goldenberg, aren't you. Go fuck yourself.
Name:
Anonymous2013-06-24 15:08
>>135 Java is different from C#
Java and C# are just DSLs on top of JVM and CLR, which are pretty much the same.
Name:
Anonymous2013-06-24 15:09
>>135 Wait... you're Goldenberg, aren't you. Go fuck yourself.
No it is you who are a Jew, because you defend Jewish language Haskell and harmful features, like lazy evaluation and algebraic types.
>>135
In LISP, things you'd expect to be part of the language can just be libraries. Once you have defmacro and vector, you are ready to implement CLOS.
Name:
Anonymous2013-06-25 5:04
>>139
What a piece of shit. Instead of creating solutions and making money you have to implement a language yourself because lishp is a good-for-nothing pile of shit-bricks you can build a shithouse from.
Name:
Anonymous2013-06-25 5:07
>>140
That's scheme. CL has more in it than you would want to use.
Name:
Anonymous2013-06-25 5:10
>>141
CL doesn't even have concurrency and static typing.
Name:
Anonymous2013-06-25 5:14
>>142
The lack of concurrency in the standard is unfortunate. Just use sbcl. It has threads and type inference.
Name:
Anonymous2013-06-25 5:21
>>143
SBCL doesn't even have a good REPL. It can't even repeat the previous command like any linux console can.
>>144
I still think CL is shit though. It's almost impossible to separate a lisp program from the full lisp system. Not every lisp implementation is an interpreter, but every one can do compilation of some sort at run time. It's not easy to ship a binary that doesn't contain a lisp compiler. Trying to make CL suitable for real work is kicking a dead horse. It's probably possible to get it there, but not worth the effort considering the alternatives that are available now.