Thursday, 11 March 2010


Mostly these days I suppose I'm known for MathML and XSLT stuff. Before that I used TeX rather a lot (I must check one of these days whether I have yet posted as often to xsl-list as I have to comp.text.tex). Around the time I started to use TeX (1987 or thereabouts) I was mainly programming in the functional programming language Standard ML. I haven't done so much with ML since, but the ML family of languages have been having something of a renaissance recently, especially with Microsoft's F# language.

So I've been returning to my roots and doing a bit of functional programming over the last few days, writing an article hosted at NAG on calling the NAG library from F#.


Dimitre Novatchev said...

Hi David, This is awesome.

Please, share some interesting problems you're solving with F#.

Long ago I intended starting with F#, but I could also use Haskell, or why not FXSL? :)

Dimitre Novatchev said...

Have you had a look at the problems at Project Euler? Surely nice problems to solve using a functional language.

David Carlisle said...

Mostly people are looking to the .Net interfaces to the NAG library (which is a Compiled Fortran DLL with ~1700 routines covering linear algebra, local and global optimisation, statistics, differential equations, etc...) because they are using .net languages, (or excel) for the bulk of an application infrastructure but need to call out to some accurate numerics.

The F# comments on the beta test have mainly come from financial institutions so far, but I can't really say here what kind of problems they were solving.