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

Re: implementation categories, exact rationals




On Oct 17, 2005, at 4:59 PM, Aubrey Jaffer wrote:

 | From: Michael Sperber <sperber@xxxxxxxxxxxxxxxxxxxxxxxxxxx>
 | Date: Sun, 16 Oct 2005 21:44:10 -1000
 |
 | Aubrey Jaffer <agj@xxxxxxxxxxxx> writes:
 |
 | > What is the rationale for mandating exact rationals?
 |
 | This (from the SRFI document):
 |
 | > Under R5RS, it is hard to write programs whose arithmetic is
 | > portable across the above categories, and it is unnecessarily
 | > difficult even to write programs whose arithmetic is portable
 | > between different implementations in the same category.
 |
 | > The portability problems can most easily be solved by requiring
 | > all implementations to support the full numeric tower.

Easy for who?

Implementing exact non-integers for SCM and Guile would take a lot of
work.

Implementing exact rationals seemed nearly trivial after the work needed for a "good" bignum implementation. (104 lines, including comments and whitespace, for rational +, -, *, /, versus > 2000 lines for a reasonable set of bignum operations.)

Brad