>>>>> "Richard" == Richard Fateman <fateman@cs.berkeley.edu> writes:
Richard> In general, if a function has been implemented in lisp
Richard> in essentially the same generality as in maxima, I think
Richard> we should just use the lisp version.
Richard> 1. The Maxima/Macsyma version was probably specified in
Richard> 1968 or so, and probably made sense at the time, but
Richard> 2. The common lisp ansi definition is probably better
Richard> thought out, and should be used unless it really really
Richard> breaks something.
One disadvantage of using Lisp's random is that every implementation
may be different so you get irreproducible results if you use
different implementations.
I think it's useful to use exactly the same generator on all
platforms. The question is that if we do use the same one will it be
fast enough and good enough for everyone?
Ray