Re: A Remarkble Computation Program History

Penny Anderson wrote:

>Carl de Boor confirmed that he indeed was responsible for the original
>1960's FORTRAN code as he invented bicubic splines while at GM
Research at
>that time.

>Nice to know the power of MATLAB enabled him, decades later, to give
this
>user such a succinct and powerful solution in the Splines Toolbox.

If you have a working Fortran 77 subroutine, you can create an
interface in Fortran 90 as simple as that in MATLAB by using features
such as assumed shape arrays (array sizes do not need to be passed) and
optional arguments.

For example, a call to the LAPACK Fortran 77 library such as

call SGELSD(M, N, NRHS, A, LDA, B, LDB, S, RCOND, RANK, WORK,LWORK,
IWORK, INFO)

can be replaced with just

CALL LA_GELSD (A, B)

in LAPACK95, a Fortran 95 interface for LAPACK.

De Boor's Fortran spline code is freely available at
http://www.cs.wisc.edu/~deboor/pgs/ , and there is also public domain
Fortran code for splines at http://www.netlib.org/dierckx/ . You can
create simpler interfaces for any of these codes in about an hour. Why
commit yourself to expensive, proprietary, platform-specific solutions
that may run slower and don't even produce stand-alone executables?