Effectiveness of industry test suites on advanced structural code coverage criteria

When engineers verify software products they need at the same time some guidance as to verification objectives and an idea of when it is reasonable to stop the verification activity. One such guidance comes under the form of criteria which are predominantly functional (typically, one exercises the software in ways it is expected to be used) and structural (typically, one ensures that the code is being exercised). This project focuses on the latter. A typical stopping criterion for a verification activity is to require that all the statements of the source code of the software be exercised at least once through tests; this makes sense since faults are inadvertently committed in the code. Another typical criterion is to require that alternative executions in the code be exercised. These are simple criteria and literature discusses several more advanced criteria. The more advanced criteria are seldom used because tool support is hard to find and because they tend to be more expensive to satisfy. The project is to assess how existing test suites created by engineers with the industry partner, which are meant to satisfy the simpler criteria, fare with those more advanced criteria.

Faculty Supervisor:

Yvan Labiche

Student:

Partner:

Cisco Systems

Discipline:

Engineering

Sector:

Information and cultural industries; Manufacturing

University:

Carleton University

Program: