BLAS in C

[David Hough]

The proposed BLAS in C are a pretty straightforward translation from
Fortran, and probably would not generate a lot of controversy.  I'd prefer that
the names of BLAS in C not overlap the names of BLAS in Fortran, since one
can imagine one application program exploiting diverse libraries which were
implemented in C and Fortran.

Some of the array syntax proposals that have been discussed
in NCEG might permit a shorter calling sequences.  Whether this
is a good thing to exploit may be a matter of opinion - there's a lot
to be said for a plain vanilla
interface that works like standard C and standard Fortran-77.

NCEG is primarily composed of compiler people rather than mathematical
software people.  Somebody could attend NCEG meetings
to champion these definitions as a part of the NCEG report, but a faster
route to acceptance might be to publish the definitions in SIGNUM as you
suggest, or even faster to contribute definitions and implementations into
the public domain, at least for the C-BLAS-1.