four issues for R5RS meeting at Xerox PARC

[Chris Hanson <cph@martigny.ai.mit.edu>]

   Date: Tue, 19 May 92 00:35:09 -0400
   From: "Aubrey Jaffer" <jaffer@martigny.ai.mit.edu>

   SCM used to signal an error for (/ 5 3) which caused no end of
   complaints from users.  In a system without rational numbers the only
   other possibility I read R4RS as allowing is an integer.

Regardless of the wording, which apparently does not say what it
should, there was never any intention to allow an implementation to
return exact results that are wrong.  Whether your users complain or
not is irrelevant -- the whole purpose of exactness is to provide a
guarantee that when you are given an exact result you can have
confidence that it is the correct answer.  In this case, `1' is an
incorrect answer.  If you want to return the truncated integer, it
must be returned as an inexact integer, to indicate that precision was
lost.

Obviously that paragraph needs rewriting to clarify this point.