More seriously, the influence of Haskell over F# (and even Python) is undoubted, but do you really think Haskell influenced Java Generics? (IMHO they were more inspired from C++ templates) (That is a question, not an assertion).

Phil Wadler had a hand in designing both Haskell and Java Generics I believe. Regards, Malcolm

On Wed, Jun 8, 2011 at 2:00 AM, Malcolm Wallace wrote:

...

More seriously, the influence of Haskell over F# (and even Python) is undoubted, but do you really think Haskell influenced Java Generics? (IMHO they were more inspired from C++ templates) (That is a question, not an assertion).

Phil Wadler had a hand in designing both Haskell and Java Generics I believe.

As far as I understand, Java/C# Generics and C++ templates "are" merely keywords around what we Haskellers call "parametric polymorphism". In other words, our type language is rich enough to express types like: stringConcat :: [String] -> String and concat :: Monoid a => [a] -> a using the "same" typing language, whereas the C-style languages require annotations. (You can probably guess which I prefer. I don't need keywords to tell me what the code describes, then the code describes it so clearly) I can't find any literature that specifically credits functional languages for the feature, but Bjarne Stoustrup himself acknowledges that functional programmers would tend to find template metaprogramming easier than others. He was probably aware of functional implementations (Haskell? Miranda? ML?) when he said that. I don't see the connection between Haskell's typeful programming and Python. List comprehensions are set-builder-notation-like syntactic sugar for lists. I didn't use them in Python and I don't use them in Haskell.

Were templates an original feature of C++ or did they appear in a revision of the langage ? Because C++ appeared in 1982 and Haskell in 1990. But of course, the ML languages are older, so maybe Bjarne Stoustrup took inspiration from them. (Anyway, functional features picked up in the mainstream imperative languages usually end up being just details that doesn't save much code typing and that are powerless when compared with the original functional features). I cannot help smiling when I hear Java/C#/Python developers telling about the "wondeful features" of their languages ;) (- We have generators! We can _yield_ values! [1] - Yeah, so nice... We have simple lists... with lazy evaluation) (- We have type generics! - Sigh...) Had once a microsoft representative that came to my school to hold a lecture about C#/LINQ. Told us during fifteen minutes about their python-like "yield" feature. Yay.... And to say some of my teachers were in awe [2]. [1] Amusing fact: this was me, a few years ago. [2] Well, of course, with a Java background. 2011/6/8 Alexander Solla

On Wed, Jun 8, 2011 at 2:00 AM, Malcolm Wallace wrote:

...

...

More seriously, the influence of Haskell over F# (and even Python) is undoubted, but do you really think Haskell influenced Java Generics? (IMHO they were more inspired from C++ templates) (That is a question, not an assertion).

Phil Wadler had a hand in designing both Haskell and Java Generics I believe.

As far as I understand, Java/C# Generics and C++ templates "are" merely keywords around what we Haskellers call "parametric polymorphism". In other words, our type language is rich enough to express types like:

stringConcat :: [String] -> String

and

concat :: Monoid a => [a] -> a

using the "same" typing language, whereas the C-style languages require annotations. (You can probably guess which I prefer. I don't need keywords to tell me what the code describes, then the code describes it so clearly)

I can't find any literature that specifically credits functional languages for the feature, but Bjarne Stoustrup himself acknowledges that functional programmers would tend to find template metaprogramming easier than others. He was probably aware of functional implementations (Haskell? Miranda? ML?) when he said that.

I don't see the connection between Haskell's typeful programming and Python. List comprehensions are set-builder-notation-like syntactic sugar for lists. I didn't use them in Python and I don't use them in Haskell.

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

2011/6/9 Stephen Tetley :

On 9 June 2011 09:02, Yves Parès wrote:

...

Were templates an original feature of C++ or did they appear in a revision of the langage ? Because C++ appeared in 1982 and Haskell in 1990.

Templates were a later addition to C++. There is a strong tradition of generics in OO and related languages that seems rather unrelated to typed functional programming - ADA, Eiffel and particularly CLU.

Note that the more recent C++ concepts are related to (and inspired by?) Haskell type classes. See Bernardy et al.'s interesting paper "A comparison of c++ concepts and haskell type classes": http://portal.acm.org/citation.cfm?id=1411324

Congratulations to the Simon's of course, kudos _and_ escudos from the ACM!

Idem. The Simons, GHC and the associated research very much deserve this award. Dominique

On 9/06/2011, at 8:02 PM, Yves Parès wrote:

Were templates an original feature of C++ or did they appear in a revision of the langage ?

The latter. "C with classes" did not have multiple inheritance, exceptions, or templates. Note that C++ templates are *not* the same kind of animal as Eiffel generics or Java generics or Ada generics or Haskell parametric polymorphism. The C++ template language lets you do type-level functional programming, and different instances of a common "type constructor" may in fact have quite different internal structures. C++ templates are NOT 'merely keywords around .. parametric polymorphism', they are a far more dangerous thing.

On 10/06/2011, at 6:00 PM, Yves Parès wrote:

Oh yes, having newly been using C++ at work, I realized they were a "big something" [1] that enabled you, as it were, to do whatever unstructured unholy type trickery you want, and yes, even making classes A<B> and A<C> completely different things. (BUT! We could argue over this fact: Isn't it also one of the purposes of... Type Families? Where the TF Foo :: * -> *, can yield to datatypes Foo String and Foo Int being completely different and unrelated?)

That was when Haskell went over to the Dark Side, yes. Curious observation: Object language Type language OO (C++) functional functional (Haskell) logical It makes me wonder what comes next...