Re: [tlug] [OT] Japan and ?? credit cards

[Edward Middleton <ejdmiddleton@example.com>]

Stephen J. Turnbull wrote:
> Ai Ling writes:
>  > Please excuse what may be a very stupid question, and I'm sure Matlab is a
>  > very fine program from my limited experience with it in college (math
>  > minor), but... how on earth would one use Matlab in an
>  > _accounting/financial_ environment?
>
> VAR, for one.  Whether you parse that as "vector autoregression" (in
> English, "we can predict that tomorrow will be basically like today,
> only different, in N dimensions") or as "value at risk" (which heavily
> uses the other sort of VAR).  However, normally that's only for
> prototyping.  Production analytical software is (still) mostly written
> in FORTRAN (although frontends and communications software aren't :-).
> There was a request for help putting together a conformant FORTRAN 77
> environment on the DJGPP list just yesterday (though the poster did
> preface it with "for reasons that are obscure and not relevant to the
> discussion..." :-).
>   

Standard FORTRAN77 doesn't support dynamic memory allocation, i.e. you
have to know the size of all matrix when you write the program or use an
implementation spicific extension.  At university I used FORTRAN90 for
numerical modeling in solid state physics[1], and Matlab for DSP,
Electromagnetics, numerical modeling.  FORTRANS major advantage is
speed, but depending on what your doing it is debatable whether it will
be any faster then Matlab, and if you can afford Matlab you can afford a
computer/supercomputer to run it on anyway.  Matlabs major advantage is
that it comes with optimized libraries for just about anything.

Edward

1. http://flex.phys.tohoku.ac.jp/riron/ronbun/u99edwa.pdf