If Lisp programs are written as syntax trees

then why isn't it called Trep - TREe Processor?

because ihbt and trees are just LoLs anyway

epic lols

syntax tree programming: STREP

The biggest problem of Lisp is lack of a decent type system.

>>5
Types = no thanks

we write our code without any unsafeCoerce.

Keep your snake oil away.

>>2,3
BACK TO /g/,``PLEASE''!

>>6
I don't care how you write your code. I want a decent type system.

>>6
Surely you are unsafePerformIO, but are you Abelson unsafePerformIO?

>>8
you thinnk you want it

but if you had one you would just abuse it

>>10
I don't abuse it in Haskell, why would I abuse it in Lisp?

So it turns out that the benefits of a real type system are obvious even to Pythonists:

I don’t have to waste my time on trivial bugs.
The bugs I do encounter are generally more meaningful and lead me to understanding the problem more. Note, meaningful doesn’t always mean harder.
Python taught me not to worry about speed that much. Haskell agreed with that but let me have it anyway.
Refactoring is a [i]breeze[/b]. In Python, I always had a nagging feeling that I forgot to change some small part of your code that will be important later.
Easy to scale code up to using many cores.

http://mechanical-elephant.com/thoughts/2015-04-20-becoming-productive-in-haskell/

So, some of the practical wins of static typing:
1) catching the majority of bugs;
2) thus freeing the programmer's mind to think more about the nontrivial, logical bugs or deficiencies;
3) performance - no runtime type-tags or type-checking (unless you want it, in which case you can have it with Typeable);
4) ease of refactoring - static typing improves the speed of development via super-fast reporting of errors.

Also there's another point which the author didn't mention, but it still exists:
5) documentation - types are a very concise and succinct form of it, reducing the amounts of normal documentation.

Still want to be unityped?

Oh, I forgot about safety, blast it. Add that as point

6) purity: the compiler can ensure that there's no IO in your STM code, making parallelism super-easy and correct.

Lisp was designed by someone with a fetish for linked lists.

TRIPS instead of LISP

>>7
are you the kind of idiot that doesn't know what a list-of-lists is?

>>16
yeah and what if I am, punk?

>>17
http://en.wikipedia.org/wiki/S-expression

The devil's in the lols.

>>16
A list-of-lists is just a list

>>20
In a shitty unityped language it is, yeah. In a multi-typed language, on the other hand, the compiler can actually distinguish things and help write and refactor.

Bow down and pay tribute to my glorious dub. Bask in their holy glory..