Michel Delaunay

Testability analysis for software components

In this paper, we propose to use the static single assignment form, which was originally proposed for code optimization in compilation techniques, in order to transform software components into a data-flow representation. Thus, hardware testability concepts can be used to analyze the testability of components that are described by C or Ada programs. Such a testability analysis helps designers during the specification phases of their components and testers during the testing phases to evaluate and eventually to modify the design.

Towards Mutation Analysis for Lustre Programs

Mutation analysis is usually used to provide indication of the fault detection ability of a test set. It is mainly used for unit testing evaluation. This paper describes mutation analysis principles and their adaptation to the Lustre programming language. Alien-V, a mutation tool for Lustre is presented. Lesar model-checker is used for eliminating equivalent mutant. A first experimentation to evaluate Lutess testing tool is summarized.

Testability analysis applied to embedded data-flow software

Testability is an important quality factor of software, particularly embedded data-flow software such as avionics software. A lack of testability of such software can badly affect test costs and software dependability. Testability analysis can be used to identify parts of software which are difficult for testing. In this paper, we propose the use of the static single assignment (SSA) form to transform source code generated from data-flow designs into a data-flow representation, and then we describe some algorithms to automatically translate the SSA form into a testability model. Thus, some metrics can be applied to the testability model in order to locate the software parts which induce a weakness of the testability.

Testability Analysis of Data-Flow Software

This paper is about testability analysis for data-flow software. We describe an application of the SATAN method, which allows testability of data-flow designs to be measured, to analyze testability of the source code of critical data-flow software, such as avionics software. We first propose the transformation of the source code generated from data-flow designs into the Static Single Assignment (SSA) form; then we describe the algorithm to automatically translate the SSA form into a testability model. Thus, analyzing the testability model can allow the detection of the software parts which induce a testability weakness.

Mutation analysis for Lustre programs: Fault model description and validation

Mutation analysis is usually used to provide an indication of the fault detection ability of a test set. It is mainly used for unit testing evaluation, but has also been extended for integration testing evaluation. This paper describes adaptation of mutation analysis to the Lustre programming language, including both unit and integration testing. This paper focuses on the fault model, which has been extended since our previous works. Validation of the fault model is presented.

Mutation Analysis for Reactive System Environment Properties

Reactive systems used in safety-critical domains demand high level of confidence. The development of these systems, which are submitted to several normative recommendations, is complex and expensive. Reactive systems can be developed by using the data-flow approach: many languages support this approach such as MATLAB/SlMULINK, LUSTRE/SCADE. This paper concentrates on the LUSTRE/SCADE language, especially the description of reactive system environment properties in this language. The description of environment properties, which is important for the validation (the proof and the test) of reactive systems, is not easy. Hence, we would like to use the mutation technique to consolidate this difficult task: we use the LESAR model-checking tool to detect equivalent mutants and some test case generators such as GATEL or LUTESS tools to kill non-equivalent mutants.

Testing criteria for data flow software

We propose the use of accessibility measures in some testing strategies to specify testing objectives based on a functional model. The functional model, which is founded on the information transfer within software, was used with success to analyze testability for data-flow software. The testing strategies based on this model allow specification of testing objectives in relation to faults diagnostic, i.e. they allow not only faults to be detected but also to be located in the software. The approach is applied on a dataflow design provided by THALES Avionics to specify testing objectives.

Analysis of Testability Metrics for Lustre/Scade Programs

Testing is a validation process carried out to find errors in a system. Testability metrics aims at identifying parts of a design/code that are difficult to test. In this article, we focus on two testability metrics defined for systems written in Lustre/Scade. An intuitive interpretation was proposed for these metrics. The aims of the work described here is to check whether this intuitive interpretation can be consolidated with factual evidences.

Automatic testability analysis for data-flow designs of reactive systems

Reactive real-time systems require a very high level of confidence. The validation of these systems is very important since it ensures the confidence in these systems. Two main activities of the validation are the proof of some parts of the system, and the testing of the system. However, testing of reactive systems is complex and expensive. This is where a testability analysis can help, by appraising the ease with which testing can be undertaken. In this paper, we propose a method to analyze automatically the testability of data-flow designs of reactive real-time systems. This method, which is based on the SATAN technology, allows testability of reactive systems to be measured very soon during the specification stage. Thus, weaknesses of the design can be detected, and the system can be improved in terms of testability.

Testability Analysis Integrated into Scicos Development Environment

While software is more and more complex, software testability becomes one of important software quality factors. Testability measures help designers and testers to improve software quality. In this paper, we concentrate on testability analysis of software designs. We propose an extension of the SATAN technology to measure the testability of Scicos models. This work is integrated into the Scicos environment to improve testability of Scicos designs.