In my investigation I have only found real world scientific problems,
such as the above, solved in the functional language SISAL.  
Because of SISAL's for-all and for-initial constructs, mapping from an
imperative algorithm for the FFT to SISAL is relatively straight-forward (one
of the goals of the SISAL project).  However, I have not seen any examples
of these types of problems solved in a ``traditional''
functional language--i.e. using recursion and/or pattern matching.

I am wondering if this is because
        a) no one has tried it, or
        b) I haven't looked in the right places, or
        c) it cannot be done, or
        d) none of the above.

I am hoping the answer is b.  But I can live with a.  
Please inform me if you have solved any of the above or
similar problems using a ``traditional'' functional language.

Thanks,
Vince.

I too am interested in scientific applications of functional languages. In
particular, I have started to prototype my speech recognition algorithms
in Haskell.

Other than the FFT, I have not seen any scientific programming examples
done in the lazy functional languages.  So far, I have written a
spectrogram program in Haskell, and also a program for building a
hierarchical representation of a speech signal using dendrograms.
However, neither of these is quite ready for public consumption yet.

If you are not particularly interested in lazy languages, you should check
the Mathematica Journal or the Mathematica reference book by Stephen
Wolfam.  Mathematica contains a functional sublanguage, and the Journal
contains examples of how to use it to solve scientific problems.  Because
it's intended for mathematics, it doesn't have the nice terse syntax of
Miranda or Haskell; but its "prelude" contains a lot of math oriented
functions in addition to the usual take, drop, etc.

David M. Goblirsch, The MITRE Corporation, McLean VA 22102

(703) 883-5450

--

David M. Goblirsch, The MITRE Corporation, McLean VA 22102

(703) 883-5450

  Nial Systems Ltd
  155 Queen St 9th Flr
  Ottawa, Ontario, Canada K2P 5C9
  613/234-4188

It is accompanied by an extensive set of user-contributed libraries of
operations that include what you are seeking. They offer the additional
benefit of being fairly readable by people who do not make their living
by programming, but by persons knowledgeable about the subject domain
after investing a little time to familiarize themselves with the notation.

Considered only as a system of notation, Nial has enormous expressive
power for scientific ideas that stimulates the discovery of further
scientific ideas. The nice thing about this notation system is that you
can also run it on a computer.

---

Jon Roland
Starflight Corporation, 1755 E Bayshore Rd #9A,
Redwood City, CA 94063-4142, 415/361-8141