Test Analysis and Design
Analysis versus design
Test analysis answers what to test: analyze the test basis, identify test conditions and test data needs, and trace them back to the basis. Test design answers how: turn test conditions into test cases, define test data and expected results, and select techniques. Candidates must keep test conditions, test cases and test procedures distinct - a frequent exam trap.
Technique selection follows the test basis and risk. Specification-based techniques: equivalence partitioning (valid and invalid partitions), boundary value analysis (two-value and three-value), decision tables (combinations of conditions and actions), state transition testing (sequences and invalid transitions), use case testing (flows), and pairwise/combinatorial testing (interaction pairs). Structure-based techniques: statement, decision and condition coverage. Experience-based techniques: error guessing, exploratory testing (chartered, time-boxed, not "ad hoc") and checklist-based testing.
Traps: testing only valid partitions; applying BVA to non-ordinal data; using a decision table when a state model is needed; treating pairwise as full coverage (it misses 3-way interactions); and claiming exploratory testing has no documentation - charters and session notes are expected.
Worked example. An age field accepts 18–65. Partitions: <18 (invalid), 18–65 (valid), >65 (invalid). Three-value BVA on the valid partition gives 17, 18, 19, 64, 65, 66; two-value gives 18, 17, 65, 66.
Rule of thumb: choose the simplest technique that covers the risk, and combine techniques rather than forcing one.