Is integration and differentiation code sepparate?
Subject: Is integration and differentiation code sepparate?
From: Stavros Macrakis
Date: Sun, 14 Nov 2010 11:46:32 -0500
On Sun, Nov 14, 2010 at 10:37, Raymond Toy <toy.raymond at gmail.com> wrote:
> But I, for one, am not particularly interested in fixing bugs that
> are created by random tests.
>
It is actually possible to find important, non-trivial bugs by random tests,
e.g. this multivariable polynomial factorization
problem<http://sourceforge.net/tracker/?func=detail&aid=696804&group_id=4933&atid=104933>;.
But it is usually a waste of effort to systematically check trivial,
unmotivated cases. It requires some understanding of the underlying
mathematics or algorithms (white-box testing) to distinguish trivial from
non-trivial cases. Usually it is the edge cases that are more likely to be
worth testing.
-s