Closed dmcclean closed 10 years ago
I think if we just went with the obvious mignitude
of the difference for the Kaucher
version we'd be okay.
That's probably true, but we haven't changed the mignitude
definition in Kaucher
yet either, it still has the previous semantics. (Minimum absolute value of endpoint, instead of minimum absolute value in the interval.)
ah that'd be a bug then. =)
Fixed.
This is what I was trying to do in the first place that led me to the issue with the old mignitude definition.
I'm not sure if it is of enough general interest to justify being here. But given the that I got it subtly wrong twice it might be.
I didn't add the same method to the empty of Kaucher variants because I wasn't sure how to handle one or both of the operands being
Empty
, or whether the directed definition had to be different somehow.