Re: motion35
Jurgen Wolff von Gudenberg wrote:
I think the motion is completely unnecessary, because
1. we NEED unbounded intervals
2. Nate's statement "you cannot proof anything for unbounded intervals is
wrong
Jurgen, this is being grossly mistated and taken out of context. My original
question in this regard was: how does one numerically find all the zeros of
a function on the domain [REALMAX,+Inf] for, say, cos(x).
3. The embedding of a family of bounded bot overflowed intervals harder to
comprehend and implement
It requires absolutely NO change to the implementation, as the paper
explains.
4. I see no problem if Nate implements a Kaucher extension of the bounded
subset of his P1788 library
5. It is impossible for us
It simply puts ubounded intervals in Level 1a. Why or how does that make
things impossible for you?
and we do not want to define P1788 for
Kaucher intervals only
This is not even a motion for Kaucher intervals. The motion simply requires
unbounded intervals and empty set are introduced in Level 1a. The unbounded
intervals may be interpreted as overflow so that midpoint operation is
defined on them, e.g.,
mid(omega([1,+Inf])) = mid([1,+OVR]) = h.
6. We have already spent too much time in order to provide an extension
point for Kaucher intervals
When and where did all this happen? (where are the motions, for example)
More than that I think that motion, if accepted, would lead to a
break-down of P1788 as a whole.
Can you explain exactly why? That seems an over-reaction to me.
Nate