Todd Ellner wrote: > > >> ..I think I prefer when they just ask for money :) > > >Really. I can't figure out what it is that they want. I have also > >received spam that is comprised of incoherent blithering and bizzarre > >conspiracy theories, but it usually involves Jesus :-) > > Reminds me of some people who were around Portland a few years > back. Their schtick was that "infinity" was really a number > I think the IEEE floating point spec defines "infinity" and "not a number" separately :-) I saw a talk long ago in which the difference mattered to an algorithm. The algorithm tried to use iterative approximation to find a zero-point of a function. The function had a zero point, but it was uncomfortably close to a point where the function was asymptopic to infinity, inducing a divide-by-zero error while trying to approximate the zero point. The (then new) IEEE floating point standard allowed the FPU to return a value instead of dumping core, so the algorithm could continue. ObSecurity: software security faults almost always occur around improper handling of error caes and boundary cases, e.g. buffer overflows, off-by-one errors, etc. Crispin -- Crispin Cowan, Ph.D. Chief Scientist, WireX http://wirex.com/~crispin/ Security Hardened Linux Distribution: http://immunix.org Available for purchase: http://wirex.com/Products/Immunix/purchase.html Just say ".Nyet"
This archive was generated by hypermail 2b30 : Wed Dec 04 2002 - 16:10:58 PST