Arnold Neumaier wrote:
The standard should be silent about how things are actually
represented, and only specify the effect of operations on
the representations (whatever they are).
Except for interchange formats -- if we want to deal with that.
(We might however decide that the only defined interchange
format for intervals is textual representation, thus avoiding
encoding issues completely. Yet a large part of the success
of 754 is that bit patterns of interchange formats HAVE been
defined, with only a few remaining issues such as NaN payloads
and, of course, Endianness issues that affect other types as
well, and are thus in a different category.)