Motion P1788/M0011.01: NO
My vote is NO.
Reasoning:
I see a use for reverse operations in constraint propagation, one
of the more wide-spread applications of interval computations.
Furthermore, I know that reverse operations are not the same
as usual operations with intersection. However, the discussion
has shown that these points have not been made totally clear yet.
Finally, it IS possible to implement reverse operations from
basic interval operations, on a software level, with arguably
lower efficiency and lack of portability.
I think further clarification of what reverse operations
is needed, for the group as a whole to decide intelligently.
Also, specific reverse operations need to be defined very clearly
and unambiguously, for an arithmetic standard that does what
we want it to do. Once we more
nearly universally understand the nature and scope of reverse
operations, we can better come to a consensus on whether or not
they need to be in the arithmetic standard, and, as applicable,
we could design the standard verbiage more clearly.
I'm looking forward to a revision of the motion.
Baker