No description available.

Data mining and its ability to handle large amounts of data and uncover hidden patterns has the potential to facilitate the comprehension and maintainability evaluation of a software system. Source code artefacts and measurement values can be used as input to data mining algorithms in order to provide insights into a system's structure or to create groups of artefacts with similar software measurements. This thesis investigates the applicability and suitability of data mining techniques to facilitate a the comprehension and maintainability evaluation of a software system's source code.

No description available.

No description available.

In analysing the evolution of Software Engineering, the scale of the components has increased, the requirements for different domains become complex and a variety of different component frameworks and their associated models have emerged. Many modern component frameworks provide enterprise level facilities and services, such as instance management, and component container support, that allow developers to apply if needed to manage scale and complexity. Although the services provided by these frameworks are common, they have different models and implementation. Accordingly, the main problem is, when developing a component based application using a component framework, the design of the components becomes tightly integrated with the framework implementation and the framework model is embedded in the component functionality, and hence reduces reusability. Another problem arose is, the designers must have in-depth knowledge of the implementation of a component framework to be able to model, design and implement the components and take advantages of the services provided. To address these problems, this research proposes the Attribute based Component Design (AbCD) approach which allows developers to model software using logical and abstract components at the specification level. The components encapsulate the provided functionality, as well as the required services, runtime requirements and interaction models using a set of attributes. These attributes are systemically derived by grouping common features and services from light weight component frameworks and heavy weight component frameworks that are available in the literature. The AbCD approach consists of the AbCD Meta-model, which is an extension of the บML meta-model, and the Component Design Guidelines (CDG) that includes core Component based Software Engineering principles to assist the modelling process for designers. To support the AbCD approach, an implementation has been developed as a set of plug-ins, called the AbCD tool suite, for Eclipse IDE. An evaluation of the AbCD approach is conducted by using the tool suite with two case studies. The first case study focuses on abstraction achieved by the AbCD approach and the second focuses on reusability of the components. The evaluation shows that the artefacts produced using the approach provide an alternative architectural view to the design and help to re-factor the design based on aspects. At the same time the evaluation process identified possible improvements in the AbCD meta-model and the tool suite constructed. This research provides a non-invasive approach for designing component based software using model driven development.

No description available.

The requirements phase of software development is an on-going problem for the software engineering community. The many disparate recommendations and best practices found in the literature make it difficult for software organisations to recognise which practices apply to their individual needs. The aim of this thesis is to pull together key solutions into a framework that allows practitioners to assess where their requirements process needs strengthening and to provide a means in which improvements can be achieved. In this thesis I show how I design, develop and validate a model of requirements engineering processes. This requirements capability maturity model (R-CMM) adheres to the characteristics of the Software Engineering Institute's Software Capability Maturity Model (SW-CMM) and is designed to take practitioners from an immature process capability through to an advanced capability. I contribute to the body of knowledge in both software process improvement and requirements engineering (RE) by providing rigorous detail of how a process maturity framework is developed to support RE practices. The model is generic and should apply to many software development organisations. The R-CMM guides users towards a view of RE that is based on goals and is problem driven. The SW-CMM framework is transformed into a simplified model that relates goals and problems to individual RE practises.

Society's increasing dependence upon software control and information process- ing provision has demanded comparable increases in software dependability. While the existing software dependability approach has resulted in significant improve- ments, its focus is heavily aimed towards achieving software dependability via redundant fault-tolerant mechanisms built into the software artifact to provide error-control in the presence of activated faults. Less emphasis appears to have been placed upon how software dependability can also be promoted through a fault-avoidance approach in the software creation process by incorporating hu- man redundancy and diversity. In this thesis, a process intervention which can potentially improve fault-avoidance is considered. This involves the setting of diverse development goals within important generic computer-based system con- texts in order to increase detection of potentially harmful assumptions which can result in subtle systemic conflicts that can undermine the dependability of the re- sultant artifact during the early development phases of requirements, specification and design. A search theoretic simulation model is progressed and developed to capture some of the important dynamics involved. The eventual outputs of the simulation model indicate that increased fault coverage and sensitivity can be ob- tained through the setting of diverse development goals during the early phases of software development.

No description available.

La multiplication des équipements informatiques augmente la consommation d'énergie de la maison numérique. Pour limiter cette tendance, les équipements de l'électronique grand public contrôlent leur consommation d'énergie individuellement. Les environnements répartis offrent de nouvelles opportunités de gestion de la consommation d'énergie des équipements. Ce travail propose de mettre en place une coordination intelligente entre les équipements afin de limiter la consommation énergétique de l'ensemble de ces équipements tout en délivrant les mêmes services. Notre approche déplace des composants logiciels d'un équipement à l'autre afin de maximiser l'efficience énergétique. Ces déplacements sont contraints par l'environnement et par les besoins des composants logiciels, i.e., l'hétérogénéité, la dynamicité et la qualité de service. L'hétérogénéité nécessite une modélisation de chaque équipement et de chaque service suivant des critères différenciant, e.g., ressources matérielles, présence de l'utilisateur. La dynamicité requiert de modifier la répartition des composants logiciels lorsqu'un événement significatif survient, e.g., apparition d'un équipement, afin de conserver l'efficience énergétique. Enfin, la mise en place de solutions moins énergivores ne doit pas impacter la qualité de service, e.g., satisfaire les besoins en ressources matérielles des composants logiciels. Nous proposons une modélisation de ces propriétés. Ce modèle est ensuite considéré par un système décisionnel autonome qui prend la décision de modifier la répartition des composants sur les équipements afin d'atteindre l'objectif d'efficience énergétique tout en satisfaisant à la qualité de service. / The proliferation of computer devices increases the energy consumption of the digital home. To limit this tendency, consumer electronic devices control their own energy consumption. Distributed environments offer new opportunities to manage the energy consumption of a set of devices. This work proposes to set an intelligent coordination between devices in order to limit the energy consumption of the set of devices while providing the same services. Our approach moves software components from one device to another in order to maximize the energy efficiency. The migrations are constrained by both the environment and the needs from software components, i.e., the heterogeneity, the dynamicity and the quality of service. Heterogeneity needs a model of each device and each service according to different criterion, e.g., hardware resources, user presence. Dynamicity needs to modify the distribution of the software components when a relevant event occurs, e.g., appearance of a device, in order to conserve energy efficiency. Finally, less energy consuming solutions must not impact the quality of service, e.g., satisfying the required hardware resources of the software components. We propose to model those properties. This model is then considered by an autonomic decision-making system. On the observation of relevant events, the system takes the decision to modify the distribution of the software components on the devices in order to reach the energy efficiency goal while satisfying to the quality of service.