No abstract

Requirements elicitation involves gathering requirements from different stakeholders. Different stakeholders often hold different views of how a system should behave, resulting in inconsistencies between their descriptions. Rigorous consistency checking methods can be effectively applied if the different views are formally specified. This is possible because of the unambiguous and precise nature of formal specification languages. However, ensuring that each formal view is self-consistent is critical before checking that the different views are inter-consistent. In this paper, an algorithm is proposed to check the self-consistency of the classes of an Object-Z specification. The proposed approach combines specification testing, model abstraction, and model checking to perform the verification.

Software merging is needed at different stages of software development to combine the artifacts created or modified by the parallel work of the different developers involved in the project. An accurate matching approach is the key to successful software merging as well as to conflicts identification. In this paper, a new matching approach for ObjectOriented formal specifications is proposed. Object-Z is used as a specification language. However, the proposed approach is meant to be applicable to a wide range of Object-Oriented software artifacts. Merging formal requirements specifications is motivated by the fact that it could help in identifying (and resolving) conflicts that will cost higher to identify (and resolve) at later stages of software development. The proposed approach incorporates heuristics for both syntactic and structural similarity. The empirical results obtained through a prototype implementation of the proposed approach were very encouraging.

Reusing program modules is one of the most direct and useful ways to develop cost effective software, especially with the availability of huge amounts of open-source projects. Reuse saves cost, increases the speed of development and improves software reliability. The quality of popular open-source packages is good as shown by various empirical studies. However, the quality of less known packages and the large number of projects developed by programming enthusiasts is unknown. Reusing them may be the source of more problems rather than being a solution to a problem. In this paper, a conceptual model is proposed to measure the reusability of program modules. A balanced set of wellestablished metrics is used to compute the most significant factors of reusability and an empirical investigation is conducted to measure the reusability of the modules of randomly selected open-source Java projects. The results obtained are thoroughly analyzed to discover the real reusability potential of open-source program code together with suggestions on how to improve this potential.

Requirements specification is a collaborative activity that involves several developers specifying the requirements elicited through several stakeholders. Operation-base merging allows combining specifications using the information available about their state as well as their evolution or change. Thus, leading to a more precise, accurate and efficient merging. Differencing specifications is a tedious, complicated, and a crucial process needed for operation-based merging of specifications resulting from collaboration. An approach for differencing Object-Oriented formal specifications is proposed in this paper. The difference is modeled as a set of primitive operations and is produced based on the matching results of specifications' elements. These matchings are calculated based on an approach employing elements' syntactic and structural similarities. The proposed differencing approach is empirically validated.