Issue: FORMAT-ROUNDING (Version 1)

[David A. Moon <Moon@STONY-BROOK.SCRC.Symbolics.COM>]

I agree with JonL's comment that IEEE round-to-nearest, rather
than round-towards-plus-infinity, should be used.  Of course IEEE
doesn't deal with decimal rounding, but the obvious analogy to
binary rounding should be employed.

I doubt that it's appropriate to say anything about this in the
standard, though, since Common Lisp leaves all other aspects of
floating-point rounding to the implementation's discretion.