You have been asked to determine a TAS for a new release of a SUT, test should be automated wherever. The new release will consist of 5 new interfaces and an amendment to 3 existing interfaces. The…

Question

Options

• AAutomate tests at both Component and System Level. Only do this automation once every
• BAutomate tests at one level only, System level. Use only the newly developed interfaces and do
• CAutomate the tests at two levels, Component and System level. Create customized hooks at
• DAutomate a test at once level, component level, Create customized interface/test hooks for this

• A
  67% (20)
• B
  10% (3)
• D
  23% (7)

Explanation

Options A and C are correct because both advocate automating at two levels - Component and System - which is essential in an incremental delivery context. With interfaces arriving across 3 sprints, component-level tests catch defects early on individual interfaces, while system-level tests validate end-to-end integration; neither level alone provides sufficient coverage. Option C's explicit mention of customized hooks/stubs is particularly important: when some interfaces haven't been delivered yet, you need test doubles to simulate them so automation can proceed without blocking.

Why B is wrong: Restricting automation to System level only sacrifices early defect detection and violates the "shift-left" principle; it also ignores the 3 amended existing interfaces, leaving regression risk uncovered.

Why D is wrong: Component level only misses integration and end-to-end validation - you won't know if the 5 new interfaces work together as a system until it's too late.

Memory tip: In incremental/agile TAS design, think "Two levels + hooks for gaps" - you always need Component and System coverage, and wherever an interface isn't delivered yet, a custom hook/stub fills the gap so automation never has to wait on the sprint.

No community discussion yet for this question.