Re: Bisection methods (was Re: Structured support...)
Nate,
On 05/25/2012 08:33 AM, Nate Hayes wrote:
John,
.
.
.
As I explain in section 3.2.2, 3.2.3 and 3.2.4 the mathematical
>definition of Interval Newton, for example, is (13), but (13) is
>not an algorithm until some exact method of choosing x \in X is
>specified, e. g., (14). Same for centered forms, B&B, etc.
In the model you are proposing, only the mathematical definitions are
defined at Level 1 and Level 2, but the corresponding algorithms are not;
according to the standard, such algorithms must return NaNs.
I don't understand this. Where in the standard (and do you mean
the current P-1788 draft?) does it say that a particular algorithm
must return a "NaN"? Are you talking about an algorithm implementation
in which we use the midpoint operator without checking whether or not
the interval is finite? If so, we might get back to the question of
exception handling, which I point out without much decor.
Baker
--
---------------------------------------------------------------
Ralph Baker Kearfott, rbk@xxxxxxxxxxxxx (337) 482-5346 (fax)
(337) 482-5270 (work) (337) 993-1827 (home)
URL: http://interval.louisiana.edu/kearfott.html
Department of Mathematics, University of Louisiana at Lafayette
(Room 217 Maxim D. Doucet Hall, 1403 Johnston Street)
Box 4-1010, Lafayette, LA 70504-1010, USA
---------------------------------------------------------------