[e-lang] An attack on a mint

[Karp, Alan H]

Bill Frantz wrote:
>
> I have used random number driven test drivers for some testing in
> the past. Random number testing has turned up problems which I
> would not have found (until after the code went production) by
> normal techniques. I like the technique.
>
That's what Intel used to check its divide hardware :(  Actually, using random data is a good technique, but Kahan has shown the importance of including specific boundary cases as well.  Hence, the n+1>n test for integer overflow.

________________________
Alan Karp
Principal Scientist
Virus Safe Computing Initiative
Hewlett-Packard Laboratories
1501 Page Mill Road
Palo Alto, CA 94304
(650) 857-3967, fax (650) 857-7029
http://www.hpl.hp.com/personal/Alan_Karp