When testing should be done:
(1) at initial implementation
(2) periodically during the year, depending on the risk and likelihood of an error
(3) following system modifications and/or upgrades
Failures in software testing may involve:
(1) failing to do any type of testing at all
(2) failing to test with a proper plan
(3) failing to test enough to detect low frequency errors
(4) failing to test all features
(5) failing to test constraints on data that can be entered to detect illegitimate values
(6) failing to test input across the entire spectrum of legitimate values
(7) failing to test after an upgrade or modification
(8) failing to track and communicate problems, especially for recurrence (broken fix)
(9) failing to test modifications, corrections or changes made to data