MORE INFORMATION ON combination.f
In a case where a program contains several errors, it is important to take into
consideration the combined effects of the errors on the reliability of the testing strategy.
This sample code characterizes a set of conditions under which the reliabilty of Path analysis testing,
(or P-testing) for single errors (demonstrated in the programs: Computation.f, and Domain.f) is preserved
when errors occur in combination.
This theory is defined by William E. Howden thus:
Suppose E = {E1 , E2 , ....Ek } is a set of errors in a program P.
Let PEi be the program which contains only the error Ei and not
the other errors (ie, Ei is the difference between P and a given
program P*). Let PE = P. An error Ei in E is independent in E if
for all x in the domain of P,
An error Ei in a program is independent relative to a set of errors
E if the introduction of the other errors into the program does not
"correct" any of the incorrect output caused by the error Ei.
Example:
PE1 is the program containing the missing assignment SUM = X.
PE2 contains the transfer statement with the missing DABS function and
PE3 is the assignment with the missing parentheses.
PE2 is the program Domain.f
Click here for the
source code