A Flavors nettle
John Pryce wrote:
> IMO, either
>
> (A) A flavor always has Entire as an interval, and Entire belongs to every
> interval type (as in the set-based flavor). If so, neither classical
> Moore arithmetic, nor Kaucher arithmetic as originally put forward by
> Nate Hayes, can be a flavor. Or
>
> (B) We permit a flavor not to contain Entire, but we define a new mandatory
> exception called, say, IntvlOverflow. This is signaled at Level 2 when
> the Level 1 result of an operation "overflows", i.e. is too large to be
> enclosed in any interval of the destination type T.
Now that we have pretty much settled on requiring a few named exceptions,
I suggest (B), because the whole point of the "Flavors" concept was to
allow Kaucher, and perhaps even Classical. And since we clearly state
that Empty and unbounded intervals are not "common", it seems natural
that Entire is not only not common, but may be missing entirely from a
flavor (sorry for the pun).
I would not even say "should include Entire", because a certain flavor
either does or does not have this concept, and suggesting that it be
shoved in just to satisfy a "should" may be a bad idea, as that could
lead to unforeseen inconsistencies in that flavor.
The exception is needed however because under no circumstances must we
allow containment to be violated (in a commonly observable manner).
I would also suggest the name ContainmentViolation for the exception,
in case there are situations where "overflow" might not be meaningful.
Michel.
P.S. One more suggestion: we have to make it clear that a flavor's
"Contains" relation *must* be a partial order relation.
---Sent: 2013-11-20 14:54:14 UTC