The main thing my team implemented in Haskell was a simulator simulator that took the first two rounds worth of port information and used that to back into the velocity for each of the satellites. This meant we really only had to appeal to the actual VM implementation for canonical scoring and to validate results, which was kind of nice. We implemented the VM itself independently in python and C to cross-check ourselves and I later put together a small VM->C compiler once we knew we had the implementation right.
-Edward Kmett
 
2009/6/30 Eugene Kirpichov <ekirpichov@gmail.com>
We implemented the VM by writing a smallish compiler to C for it in
Haskell. It ran *damn* fast, but we couldn't get rid of some bug that
did not let us run the 4rth task at all, although the others worked
fine :(

2009/6/30 Ahn, Ki Yung <kyagrd@gmail.com>:
> John Meacham 쓴 글:
>>
>> I implemented the VM in C, it was pretty obviously geared towards
>> such an implementation and it took all of an hour. Then I interfaced
>> with it via the FFI. Why use just one language when you can use two? :)
>
> You could also have used Data.Binary.  That's what I did.
>
>> I wasn't able to make any time on sunday though so didn't end up
>> submitting a final entry which is too bad. this was an interesting one.
>>
>>        John
>
>
> _______________________________________________
> Haskell-Cafe mailing list
> Haskell-Cafe@haskell.org
> http://www.haskell.org/mailman/listinfo/haskell-cafe
>



--
Eugene Kirpichov
Web IR developer, market.yandex.ru
_______________________________________________
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe