Wed Mar 29 01:50:00 GMT 2000
It's unlikely in practice but I believe, in principle, the
representation of zero could be different in an implementation-
dependent extended floating point format, such as long double.
Thomas Walter writes:
> Is there a special reason to not use the systems 'calloc()' but
> 'malloc()' followed by a loop over all elements to set them to a
> initial value of zero?
More information about the Gsl-discuss