Re: F# fsharp

> Sorry, I was confusing Haskell with OCaml. Which still isn't *widely* used
> in the financial industry, like Java, say, but at least is used in anger in
> some places :-)

I guess this depends on your definition of 'widely'. I don't believe anyone is arguing that Haskell is widespread across the financial sector, but it definitely is used. I can vouch for at least one case as I've met several of the guys working with Haskell at Standard Chartered, have some idea of what they are doing and it all sounds very interesting and productive. Definitely not a fictional story.

The critical difference between F# and Haskell is Haskell's pure functional approach. In a nutshell this adds constraints to the way you code but a) can be a big help in composing functions and controlling side effects, and by implication b) makes it easier to write parallel-friendly code. As I understand it, F# could not adopt this approach as they have a requirement to be able to call anything in .net at any point - essentially choosing much tighter coupling with .net over a pure functional approach. 

This is just another way of saying they are different languages with different goals. I don't believe games really value interop with .net to the same degree, and I'd argue games have quite constrained IO - they often benefit from appearing more like a functional pipeline - so I personally think Haskell is a better fit in this regard.

I also wrote a blog piece after the stackoverflow post, which focused on different reasons Haskell is not so widely used. It's rather less personal than Jon's (Jon of Flying Frog fame, not jw).

http://www.palgorithm.co.uk/2010/05/why-isnt-haskell-popular-in-industry/


> Too small a number of people
> fighting over intellectual and financial scraps.  You can't really
> talk to these people seriously about money, productization, or
> usability, as they're simply not motivated that way.  Personally I
> think the commercial merit of F# is bloody obvious compared to any of
> the other FP competitors.  The only thing actually worth discussing
> IMO is whether F# can replace Lua, Python, Ruby, C#, etc.

I do feel this is rather unfair, although I'm not totally sure who 'these people' are? I have not found the FP community to be against wider adoption of Haskell in industry, quite the reverse. In particular, I've found all the GHC guys to be very keen to both promote and understand the problems faced by people interested in adopting Haskell outside academia. This isn't surprising - industry adoption of your research is perhaps the best validation you could get.

Cheers,
Sam

_______________________________________________
Sweng-Gamedev mailing list
[email protected]
http://lists.midnightryder.com/listinfo.cgi/sweng-gamedev-midnightryder.com