Re: Software Reliability



Date view Thread view Subject view Author view

Mark T Miller (mtmiller(at)crnotes.cca.rockwell.com)
Mon, 30 Jun 1997 12:51:34 -0500


Jon Davies wrote (in response to others0: > I can't see any way that software failures would be dependent on time. > software failures are (usually) systematic - and so are dependent on a > particular piece of (faulty) functionality being exercised. I agree; this matches my experience, and the characteristics of safety-critical avionics software that I've seen. >> What are the opinions about Alpha and Beta tests. > Personally, I think very little of alpha and beta testing in the context > of >>safety critical systems. I agree with this statement, too. It's the boundary conditions, expecially the unexpected ones, that cause the problems, and these boundary conditions can go untested for a long time unless I explicitely write tests based on the requirements and design of the software and system. mTm


Date view Thread view Subject view Author view