[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: arithmetic issues

This page is part of the web mail archives of SRFI 77 from before July 7th, 2015. The new archives for SRFI 77 contain all messages, not just those from before July 7th, 2015.

> > type declarations cannot address the
> > portability and predictability issues unless implementations
> > are required to interpret those declarations in a consistent
> > way.
> By "predictability", do the authors of the SRFI mean that inexact
> arithmetic gives the same result everywhere, that (/ 1 2) gives
> the same result everywhere, or something else?
> Regards,
> Alan

We do not mean that inexact arithmetic will give the same
result everywhere.  We do mean that (/ 1 2) and similar
computations will give the same result everywhere.

"Predictability" is my word for what I think Mike means
by "transparency".  Predictability also has to do with
making it easy to ensure that a computation involves
only exact or only inexact numbers, which is the main
motivation for the exact and inexact operations of this
SRFI, and is one of the main motivations for the flonum

> I would say that R6RS has to provide arithmetic on exact
> integers that are large enough to serve as indexes and
> lengths. So, either R6RS must require that fixnums are
> large enough or it must provide routines to perform
> arithmetic on integers that might be larger than fixnums.

That is a good point, and I am in complete agreement.