- Sufficient up-front money is not allocated to test software early in the design phase, to prevent design and coding errors from being discovered after deployment.
- Testability is not built into the software design.
- Software test group members are not active participants in design reviews.
- Software testers are not independent of developers and programmers.
- A closed-loop failure reporting system is not used during software test.
- Stress testing and worst case testing are not used to ensure adequate design margins.
- Security requirements are not considered during software test.
- There are not frequent, separate inspections ("walk-throughs") conducted as informal reviews on each software module.
- Reports from walk-throughs at the detailed design and coding level do not exist.
- Software reviews are conducted without government participation.
- Test Readiness Reviews are not used to ensure good software test planning.
Cost Of Correcting Software Errors
Based on data from four major subcontractors, the figure listed below shows the cost of correcting a software error during operation is 100 times the cost of correcting an error during design. Only effective software testing can identify errors at an early stage to keep costs down, to prevent unnecessary schedule slippages, but most importantly to ensure that a reliable product will perform its mission in the field.