Comments on: Static Typing and Functional Languages

By: Olle Kullberg

Olle Kullberg — Fri, 11 Jun 2010 10:26:26 +0000

Mehrdad “C# has the type inference” <- It is not true. You must mean F#

By: Tony Morris

Tony Morris — Sat, 20 Mar 2010 12:15:39 +0000

The problems in Functional Java are not because of static typing. They are because of Java. I recommend taking a look at a practical static typing implementation some time.

By: Mehrdad

Mehrdad — Tue, 15 Sep 2009 02:54:21 +0000

Have you tried C# or F#? C# has the type inference and makes it a lot simpler to do the kind of tricks you were talking about. I think it is a good balance between both functional world and traditional OOP.
I think that is needed for transition from structural and OOP to functional. Very much the same way that C++ was a transition from structural to OOP.

By: Jason Kikel

Jason Kikel — Thu, 03 Sep 2009 12:43:27 +0000

Link messed up for some reason.. The magic is called Hindley-Milner type inference. It freakin' saves the day.

By: Michael Greenberg

Michael Greenberg — Thu, 03 Sep 2009 02:53:15 +0000

http://en.wikipedia.org/wiki/Type_inference#Hindley.E2.80.93Milner_type_inference_algorithm

Type inference is a very old idea. The Damas-Hindley-Milner algorithm was a big step in that it supported a stylized form of System F, which allows it to account for precisely the example you give.

If you’re interested in learning more, check out Benjamin Pierce’s book, Types and Programming Languages.

By: Robin Bate Boerop

Robin Bate Boerop — Thu, 03 Sep 2009 01:26:19 +0000

What do you think of Haskell? How would that change your conclusion?

By: Fred Blasdel

Fred Blasdel — Thu, 03 Sep 2009 00:22:03 +0000

Hindley-Milner type inference is not advanced, and it did not take a long time to develop!

The original algorithm was first published in 1969, and the first ML implementation was done by Luca Cardelli in 1973 (the same year as K&R C!). I’m pretty sure he was an undergraduate at the time!

By: Paul

Paul — Wed, 02 Sep 2009 23:44:41 +0000

Whoops, by “it” I meant “Scala”

By: Paul

Paul — Wed, 02 Sep 2009 23:44:16 +0000

Forgive me if I’m wrong, but I was under the impression that it ran on the JVM, and in fact would compile down to Java files…

By: Brad Clow

Brad Clow — Wed, 02 Sep 2009 23:10:01 +0000

It is interesting to note that type inference appeared in the 1970s, well before Java existed. Functional Java ‘looks’ so ugly, because Java is not naturally suited to expressing these abstractions.

I suspect the statement that functional programming implies that static typing is difficult, is an error. Type system research has been motivated by a number of factors over the years, including compiler optimisations for program performance and theorem proving.