Software Quality Assurance & Testing Stack Exchange is a question and answer site for software quality control experts, automation engineers, and software testers. It only takes a minute to sign up.
Sign up to join this community
Achieving branch coverage is possible but it doesn't mean complete testing.
Achieving branch coverage is possible but it doesn't mean complete testing.
Indeed, no amount of fact-checking can achieve complete testing, because fact-checking is only one of the activities of testing, which also includes modeling, learning, experimentation, etc.
What kind of defect you can find with this technique?
Code and mutation coverage are heuristical validation methods for detecting missing automated fact-checks. So they don't find defects in the product per se because they don't spawn oracles. However, they can alert the developers that additional fact-checks may be created - and the process of creating such fact-checks will determine the defects that they can detect.
what kind of risk you could have with this technique?
As with any tool, the main risk is the misunderstanding of its goal and limit. Many teams use them for blind gatekeeping, others limit their testing to reaching some number on these metrics.
what possible defect you could miss when you use only this technique?
"Quality is value to some person", Gerald Weinberg.
Coverage methods talk only about the code that was implemented. Since quality is a subjective and dynamic aspect of someone's perception, no static and objective documentation (code) is able to correctly cover all possible aspects of quality. Additionally, since we have explicit and tacit knowledge, it's impossible to even explicitly describe the desires of a stakeholder in a particular moment.
This means that fact-checks are technically limited to detecting deviations from past and static facts we have found important/interesting. Anything outside this scope "escape" to production.
