Re: F# fsharp

On Tue, May 25, 2010 at 4:13 PM, Nisarg Kothari <[email protected]> wrote:
> I know there is a company in finance that uses OCaml for that same reason:
> they believes it gives them a competitive edge.

Yes, that has been known / done in financial markets for awhile now.
Back in the day, the reason I didn't fully swallow OCaml was because
it seemed to be *only* finance markets and academic researchers who
cared about it.  There just weren't enough dollar signs written on it
for the rest of the computer industry.  Admittedly, the dollar signs
on financial software can be pretty good, but I was trying to do
games, not Wall Street data mining.  OCaml didn't have much in the way
of support for 3D anything.  DirectX, forget about it, and anyone's
given OpenGL binding was stale in the typical open source fashion.
The OCaml guys didn't care much about float performance or support,
and also adulterated their types by grabbing an extra bit out of them
for garbage collection purposes.  This rubbed my ASM sensibilities the
wrong way, and what I really wanted was some kind of high level
programming language that would go down to pretty good assembly code.
OCaml wasn't it, and it became clear that functional programming
researchers in general didn't care about my concerns.  Several years
ago, F# was just some research lab thing, one of the plethora of
languages that Microsoft was paying people to get working with .NET.
It was simply a slower OCaml that did boring .NET stuff, and didn't
have any industrial following, so who cares?  Well, that was then and
this is now.

> Another example is Paul
> Graham. He claims that using Lisp for some server software allowed his small
> team to stay competitive with a big company that was using C for their
> product.

Yeah but he disappeared into Arc development land, a programming
language of his own design.  Nothing important has come of it, and
he's been at it for a long time.  I've been inspired by some of Paul's
essays, particularly those about being some kind of maverick hacker
programmer god who bucks the received wisdom of how to make money, and
how to work towards The Next Big Thing.  But there's a point at which
I start feeling like it's all talk and not a lot of real world
implementation.  Similarly, whatever the general or historical merits
of Lisp, in the real world of industry nobody's managed to deploy it
in a compelling way over the past decade.  Other languages have
gradually co-opted Lisp features, to the point that now, you could
write Lisp code in F# if you were so inclined.  Instead of CAR and CDR
they call it FST for "first" and SND for "second", which is definitely
easier to remember.  Sure there was an uptick in interest in Lisp over
the past several years, but Python, Ruby, and Microsoft didn't sit
still during that time.  Eventually programming language constructs
become a semi-commodity and you can't just coo forever about how great
something was 50 years ago.  Or even 15 years ago, in the case of Paul
Graham.  That was then and this is now.


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