
April 27, 2007
8:29 p.m.
Anton Sherwood wrote:
Jan Grant wrote:
. . . the space of doubles include a small number of things that aren't representable by a rational (+/- inf, for instance).
Anton Sherwood wrote:
+1/0, -1/0
David Goldsmith wrote:
Are not unique representations for inf/-inf, (if that matters).
No rational representation is unique!
Of course - as soon as I hit send I knew someone was going to shoot that back at me! :-) But for some reason (I'm speaking rhetorically, I know the reasons) they always drilled home non-zero/zero undef., not = inf. DG