[SystemSafety] Fault, Failure and Reliability Again (short)

C. Michael Holloway c.m.holloway at nasa.gov
Tue Mar 3 15:04:00 CET 2015


On 3/3/15 2:11 AM, Peter Bernard Ladkin wrote:
> I had some private discussion with someone here who claims software cannot fail.
Perhaps this person is using "fail" to mean "wear out."  Software cannot 
wear out.  That is, software does not work properly one day in response 
to a given sequence of inputs in a particular environment, and then 
cease to work properly the next day in response to the identical 
sequence of inputs in the identical environment.  If "fail" means "wear 
out" then "software cannot fail" is a coherent, correct statement.

(Note: I realize that the third sentence above is a simplification that 
may strictly apply only to deterministic software.)





More information about the systemsafety mailing list