Thread Links | Date Links | ||||
---|---|---|---|---|---|
Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
Vincent Lefevre wrote:
On 2009-09-16 18:35:54 -0700, Dan Zuras Intervals wrote:In particular, the semi infinite sets have no representation in midrad. An important subset in need of representation.That's not a problem. If you have a semi-infinite set, this means that the error bound is infinite. So, this isn't much different from the whole set R of the real numbers.
The little difference may be decisive, though, for example when checking algebraic numbers for their sign, using multiprecision arithmetic. x in 1e-10000 + [0,Inf] encodes sign information, which is lost when going to midrad. Arnold Neumaier