Software bugs and testing techniques

Types of bugs (by phase of introduction)

1. Requirements bugs: incomplete, inconsistent or wrong requirements; misunderstanding of what the customer wants.
2. Design bugs: wrong architecture, algorithm or interface; missing error handling.
3. Coding bugs: syntax errors, logic mistakes, uninitialised variables, array bounds, memory leaks, data-type errors.
4. Integration bugs: modules do not work together because of interface and data mismatch, timing, or hardware-software interaction.
5. Maintenance/regression bugs: introduced when changing or fixing the code.

Testing techniques to identify them

• Static techniques: reviews, walkthroughs, inspections and static analysis of requirements, design and code find bugs without running the program.
• Black-box (functional) testing: test the outputs against the specification, without knowing internals; uses equivalence partitioning, boundary value analysis, decision tables. Finds requirement and functional bugs.
• White-box (structural) testing: uses the code structure: statement, branch and path coverage, loop testing. Finds coding and logic bugs.
• Unit testing: each module tested separately (by the developer).
• Integration testing: modules combined step by step (top-down, bottom-up) to find interface bugs.
• System testing: the complete system against requirements: performance, stress, security, recovery.
• Acceptance testing: alpha (at developer site) and beta (at user site) tests by the customer.
• Regression testing: rerun old tests after changes to find newly introduced bugs.
• For embedded software, hardware-in-the-loop testing and simulation are also used.