[nycbug-talk] instructions for building f# on freebsd

Siraaj Khandkar siraaj at khandkar.net
Sat May 18 23:22:35 EDT 2013


On May 17, 2013, at 1:57 PM, Pete Wright <pete at nomadlogic.org> wrote:

> On 05/17/13 07:09, Justin Dearing wrote:
>> On Thu, May 16, 2013 at 1:50 PM, Pete Wright <pete at nomadlogic.org
>> <mailto:pete at nomadlogic.org>> wrote:
>> 
>> 
>> 
>>    I am new to F# - is there a good tutorial and/or good usecase of
>>    when I would want to use F#?  It looks like it is built on top of
>>    Microsoft's .NET CLR - is that a good thing?
>> 
>> 
>> I'll try to answer objectively. I've never written F# personally. I've
>> written a lot of CLR code in C# and powershell, and when forced at
>> gunpoint VB.NET <http://VB.NET>. I don't have a suggestion for a
>> tutorial. As far as when you want to use it, its an extension of OCAML,
>> and there is an option to compile pure OCAML code. Its a really
>> good language for implementing algotithims. Its not good for "plumbing
>> code." So as an example, if you were writing a daytrading system in the
>> CLR, you code to talk to the bloomberg terminal, and talk to the system
>> that executes the trade should be in C#. The code that's taking your
>> data from your bloomberg terminal and doing all kinds of crazy math to
>> determine when and what to buy and sell should be in F#.
> 
> ah ok thanks Justin!  We've been adopting "R" in my domain over here - so it sounds like F# may fit a similar role as that.  Got it.

R and F# are as different as they come: R is quick and dirty DSL for explorative statistics and is a poor fit for developing systems, while ML and thus F# is a language focused on safety, expressiveness and pragmatic interoperability, so it is excellent for developing complete systems [1] (not just math).


[1]: As in "business systems", not as in "operating systems" (for that it would also be terrible :))

-- 
Siraaj Khandkar
.o.  o.o  ..o  o..  .o.
..o  .oo  o.o  .oo  ..o
ooo  .o.  .oo  oo.  ooo






More information about the talk mailing list