Symbolic Methods in Testing

Summary

Recent breakthroughs in deductive techniques such as satisfiability modulo theories (SMT), abstract interpretation, model-checking, and interactive theorem proving, have paved the way for new and practically effective techniques in the area of software testing and analysis. It is common to these techniques that statespaces, model-elements, program-fragments or automata are represented symbolically making systems amenable to analysis that have formerly been out of reach. Several research communities apply similar techniques to attack the classical problem of state space explosion by using symbolic representation and symbolic execution: parametrized unit testing, fuzz testing, model-based testing, theoremprover based test case generation techniques, and real-time system testing. Moreover, several areas where symbolic methods are used in testing, are often considered more closely related to verification and end up in conferences specialized on those topics rather than at testing conferences. There is little synergy between the different communities although many of them use similar underlying symbolic techniques.

In the following areas, symbolic analysis techniques have recently had significant impact, both industrially as well as in academia. The following areas capture some topics of interest for the proposed seminar, assuming focus on the use of symbolic techniques in each area: Unit Testing, Symbolic Automata Theory in Testing, Model Based Testing, Fuzz Testing, Security Testing, Real-time System Testing, Theorem-Prover-based Test-Case Generation, Hybrid System Testing, and Mutation Testing.