Program

See below for a list of the presenters and download the slides

Scheduled lectures

• Andrea Polini (ISTI-CNR), Lars Frantzen (ICIS-Radboud University Nijmegen). Slides
  Modeling and Testing Service Oriented Architectures. Abstract
  Service Oriented Architecture (SOA) is the emerging paradigm for developing complex, distributed applications. One of the key promises of SOA is the interoperability among applications developed and deployed by different organizations on different distributed machines. Hence, the quality of a SOA is a complex measure, which may depend on several challenging characteristics such as higher dynamics due to run-time discovery and binding, reduced trust given the reduced control over the invoked services, and potentially even advanced concepts like context awareness and adaptability. Regarding testing SOA, current research aims at discovering which are the new opportunities and challenges in this domain. The first part of the talk will introduce the concepts and standards relevant for developing and modeling SOA, focusing on Web Services as the key technology. The second part aims at presenting interesting research directions in the area of service testing. At the same time some proposed solutions will be introduced.
• Arnaud Gotlieb (INRIA-IRISA).
  Constraint-based testing. Abstract Slides
  A new trend in program testing is to combine static and dynamic analyses through the usage of the Constraint Programming technology. Constraint-Based Testing (CBT) was introduced a few years ago in the context of mutation testing to model processes of test case generation using constraint solving techniques. Since then it has continuously evolved to cover several application areas including hardware verification, test data generation for structural testing, model-based testing, or counter-example generation in software verification. In this lecture, I will survey the principles of CBT, its main implementations and some recent advances in the field.
• Lars Frantzen (ICIS-Radboud University Nijmegen).
  Formal Testing - Theories for Practice. Abstract Slides References
  According to Glenford Myers "testing is the process of executing a program with the intent of finding errors". To define what is meant by "error", a specification of the desired behaviour of a system is mandatory. Having such a specification, further questions arise, like: "How do we generate test cases out of the specification?" "What does it mean for a system to be conform to a specification?" "What properties should a test case generation algorithm have?" "Under which hypotheses can we stop the testing and still draw conclusions about the conformance of the system?" Formal testing tries to answer such questions. The talk tries to summarise the several approaches, focusing on the main branches related to model-based testing of transition systems.
• Pascale Legall (Université d'Evry Val d'Essone - LAMI), Christophe Gaston (CEA List).
  Symbolic execution technics for conformance testing. Abstract Slides
  The content of this lecture deals with conformance testing. Conformance testing is a testing activity for which the System Under Test (SUT) can be seen as a black box, that can be controlled and observed through its interface and for which the expected outputs are described in a specification. Thus, the aim of conformance testing is to reveal SUT behaviours that do not conform to its specification SP by selecting test inputs from SP and by emitting a success/failure verdict by comparison of the SUT outputs and the outputs specified by SP. When SP is described by means of an automata-based language, (non-)conform behaviours are described in the form of sequences of inputs/outputs. In order to construct those behaviours, sequence of inputs are selected and sent to the SUT according to some test purposes and to the first outputs observed from the SUT. Verdicts are emitted by comparing SUT outputs and outputs specified by SP in relation with the test purpose given as parameter. In this talk we focus on the particular case of symbolic specifications : data handled in automata are not enumerated, but rather denoted in a first order style. Data are used both to control, by means of symbolic guards on transitions, which inputs or outputs are allowed, and to store or exchange information with state variables and messages carrying data. We focus on a particular technique, called symbolic execution, allowing to reason on symbolic data: intuitively, a symbolic path of the specification aggregates a set of concrete (or numerical) paths which share a common behaviour with respect the specification. We discuss how symbolic execution, combined with constraints solving, can be fruitfully used to compute test purpose, define selection criteria, generate test data and assign verdicts. We will illustrate on some training examples and give some hints on how symbolic execution can still be involved for extensions as refinement testing or testing of component-based testing.
• Ana Cavalli, Stephane Maag (National Institute of Telecommunications
  Security and interoperability testing of ad hoc network protocols. Abstract Slides (Ana) Slides (Stephane)
  Formal description techniques and their testing tools are rarely applied in the MANETs. The main reason is the difficulty to take into account their inherent constraints and the mobility of nodes in the test sequences generation and their execution. We present in this talk techniques to generate test scenarios allowing to test the interoperability of routing protocols (DSR). Nevertheless, the execution of these scenarios is currently an issue. Indeed there is often a gap between the dynamic topology designed in a specification and the one of a real case study. Node selfsimilarity concept will therefore be illustrated. Besides of interoperability testing, security aspects will be mentioned and techniques to detect attacks through the MANET routing protocols will be introduced.
• Rob Hierons (Brunel University), Lydie du Bousquet (Université Joseph Fourier - Laboratoire LIG).
  Software testability. Abstract
  "Testability is the relative ease and expense of revealing software faults." In his article "Design for Testability in Object-Oriented Systems", Binder has identified 6 main facets influencing testability during software development: representation, implementation, built-in-tests, test suite, test tools and test process. During this talk we will discuss on Binder's vision of testability and several other works on software testability.
• Ina Schieferdecker (TU Berlin).
  Model Transformers for Test Generation from System Models. Abstract Slides
  The early integration of test development into the system development process becomes more and more important. By doing so, design mistakes and implementation faults can be detected in an early stage of the system design and implementation process, i.e. before the newly developed system is shipped to the customer. This allows for reducing the overall development time and costs significantly. This talk reports on results in integrating testing into a model-based development process. Based on a common MOF-based infrastructure, initial test are automatically derived from the system models. Subsequently, an executable test code can be automatically generated from the test models and being applied to the final system for quality assessment.
• Marc Aiguier, Delphine Longuet (Université d'Evry Val d'Essone - LAMI).
  Testing dynamic systems from axiomatic specifications. Abstract Slides (Marc)
  This talk deals with test case selection from axiomatic specifications. The starting point of the selection process is the definition of an exhautive test set, that would have the power to prove the correctness of the program if it could be submitted in its whole. This test set being infinite, a selection phase is needed to build a test set of reasonable size to submit to the program. A first step in selection is making a partition of the exhaustive test set in order to apply the uniformity hypothesis to each subset of the partition. This partition will be done by a method called axiom unfolding. We will present here several exhaustivity results, as well as the method of axiom unfolding, in the framework of algebraic and first-order specifications.
• Ismail Berrada (Laboratoire Informatique, Image, Interaction).
  Testing Timed Systems. Abstract
  ABSTRACT :

Industry day (IUT 2 (amphi 2), Place Doyen Gosse, Grenoble)

• 10:00am : Christophe Grandpierre. (LEIRIOS). Slides
  Model-Based Testing from UML models.
• 11:00am : Nicky Williams. (CEA List LSL). Slides
  PathCrawler : automatic generation of test inputs for structural coverage of C functions.
• 2:00pm : Bruno Marre. (CEA List LSL). Slides
  GATeL : test sequences generation from LUSTRE descriptions.
• 3:00pm : Bruno Dillenseger. (Orange Labs). Slides
  The CLIF load test platform and its utilization at Orange Labs.

PhD talks

Monday July 2nd

Thursday July 5th

Friday July 6th