On 2009-04-13 09:19:34 -0400, Michel Hack wrote:
If the underlying arithmetic is not correctly-rounding, it would be
difficult to define "tightest bound" for the enclosures of primitive
IA operations, which we are considering for the standaed. In this
context I interpret "tightest" as saying that no narrower representable
result would guarantee containment.
Is the "tightest bound" required or only recommended? Having it
required would be annoying if one wants interval arithmetic on
the elementary functions, as IEEE 754-2008 only recommends
correct rounding for such functions.