[GAP] linear algebra package

[Martin J. Stift]

Some time ago I wrote:
>> I am continuing to use Ada95 in my courses on modelling of stellar 
>> atmospheres, spectral synthesis and polarisation and I have not yet 
>> found anything really important or useful for high performance object 
>> oriented scientific computing in the new features introduced by Ada2005.
and Robert Dewar suggested
> The most obvious candidate here is the linear algebra package
> which we will implement as an interface to the standard Fortran
> BLAS library. But likely you already have a library of this kind.
> It will be nice in future to have a standardized interface for
> this kind of functionality.

Now in polarisation I deal with Stokes vectors that have 4 components 
{I, Q, U, V}. In Ada.Numerics.Generic_Real_Arrays however I* * don't see 
any enumeration types but only
  *type* Real_Vector *is array* (Integer *range* <>) *of* Real'Base;
  *type* Real_Matrix *is array* (Integer *range* <>, Integer *range* <>) 
*of* Real'Base;

Do I really only have the choice between either self-made less efficient 
packages or Fortran-style stone-age indexing with {1..4} instead of {I, 
Q, U, V}? I always told my students about the advantages of enumeration 
types but now I am no longer sure about that.

Univ.Prof. Dr. Martin Stift
Institut für Astronomie
Universität Wien
Türkenschanzstr. 17
A-1180 Wien, Austria
-------------- next part --------------
An HTML attachment was scrubbed...
URL: /pipermail/gap/attachments/20070521/5b14729d/attachment.htm