FIGUEIREDO, A. L. L.; http://lattes.cnpq.br/4979118853748310; FIGUEIREDO, André Luiz Lima de.
Abstract:
Looking for reliability regarding to their systems correction, developers have, increasingly, used software testing. This activity can be defined as a set of experiments performed over an implementation, which the results are observed and analyzed. Among the existing kinds of test, this work takes special attention to a specific kind of functional test, called Formal Test. From a formal specification and through formal procedures, aformaltest method isable to,automatically, generate test cases that aim to verify conformance between specification and implementation. Mobile Agent-based systems can benefit from formal testing. Mobile Agent can be defined as an autonomous program that can
migrateamongdifferent points ofadistributed systemduring itsexecution preserving its state. Develop this kind of system is still a complex activity, due to, mainly, development processes immaturity regarding to mobility issues handling. Aiming to soften this problem, this work presents an automatic test case generation proposal to Mobile Agent-based systems. This proposal comprises presenting a system specification formalism to be used with Mobile Agents and a strategy for test case generation via tools. Moreover, aiming to take advantage of the existing UMLmodelos to this kind of system, a informal transformation from models written such language, to the formalism language used by the
presented method is proposed. After the method has been proposed, a case study was carried out addressing to show the method’s applicability. The generated test cases were analyzed with relation to its potential to be implemented and to reveal system faults. Motivated by the test cases study, and by the increasing interest atthe use of design patterns as well, a methodology foridentification oftest patterns from Mobile Agent design patterns is proposed, in addition to a test pattern template for this kind of pattern, and a set of test patterns to Mobile Agent-based systems.