fpprec was originally intended for bigfloats. It makes
sense to separately specify bigfloat PRINTING though.
1.00000000000000000000000000b0 is not necessarily what you want.
Commercial macsyma has bfprint_precision as well as bfprecision,
an alias for fpprec.
If you are printing out numbers for purposes of
FORTRAN programs, I don't see why fpprec should be
consulted, so I agree with Robert and Raymond that
full precision should be used.
I think we could specify a FORMAT for printing machine floats -- let the
user specify, for example, format: "~10e5";
RJF
>ima
>
>