Use Case modeling has been constantly gaining popularity as the technique of choice for eliciting and documenting functional requirements. The deployment of Use Case models in industry has resulted in many positive experience reports being published. The inclusion of Use Case modeling into the Unified Modeling Language (UML) (OMG 2005) has aided its widespread use in industry, especially within the object-oriented community. One of the most attractive aspects of Use Case modeling is its technical simplicity, allowing stakeholders with differing backgrounds, to have a common understanding of the requirements. This technical simplicity can be deceptive, as many modelers create models that are incorrect, inconsistent, and ambiguous and contain restrictive design decisions. In Use Case driven development processes, Use Case models are used to drive the design and test phases. While a number of techniques have been proposed to develop test cases from Use Case models, these techniques tend to suffer from two major shortcomings. The techniques are technically too complex to be effectively used by its potential users (business analysts and customers); and the inability to use these techniques in the early stages of development. This thesis describes work tackling these deficiencies. Support for developing higher quality Use Case models is achieved by developing a modeling syntax that ensures consistency when constructing Use Case models. A controlled experiment was performed to empirically evaluate the effectiveness of using the modeling syntax. In addition, a technique based on utilizing antipatterns to detect potentially defective areas in Use Case models was developed. The technique prompts modelers to revise and remedy poor design decisions, yielding superior quality models. Finally, a framework was developed, which utilizes Use Case models, to develop acceptance tests. The framework was designed to account for the technical abilities of its potential users. / Software Engineering

Thesis (Ph.D.)--University of Hawaii at Manoa, 1994. / Includes bibliographical references (leaves 259-266). / Microfiche. / xviii, 266 leaves, bound ill. 29 cm

In this dissertation we identify two problems with current integrated development environments (IDEs) and present JQuery as a tool to address these issues. The first problem is that IDE views answer low level questions and do not provide a mechanism to combine results to answer complex higher level questions. Even relatively simple questions force the developers to mentally combine results from different views. The second problem is that IDEs do not provide an easy way to create perspectives on project specific concerns such as naming conventions or annotations. Most IDEs do offer support for creating custom perspectives but the effort required to create a perspective is considerably more than the benefit a custom perspective provides. JQuery is an Eclipse plugin which generates code views using an expressive query language. We have redesigned JQuery to support a number of new user interface (UI) features and add a more flexible architecture with better support for extending the UI. To address the first problem, we have added multiple views to JQuery where each view supports drag and drop of results, selection linking, and regular expression search. These features enable a user to combine results from different views to answer more complex higher level questions. To address the second problem, we can leverage the fact that JQuery is built on an expressive query language. Through this query language we are able to define project specific concerns such as naming conventions or annotations and then create views and perspectives for these concerns through the JQuery UI.

Thesis (PhD)--University of South Australia, 1998

Thesis (MEng)--University of South Australia, 1997

No description available.

???Agent-based systems are one of the most vibrant and important areas of research and development to have emerged in information technology in the 1990s??? (Luck et al. 2003). The use of agents as a metaphor for designing and constructing software systems represents an innovative movement in the field of software engineering: ???Agent- Oriented Software Engineering (AOSE)??? (Lind 2000; Luck et al. 2003). This research contributes to the evolution of AOSE by proposing a comprehensive ontology-based methodology for the analysis and design of Multi-Agent Systems (MAS). The methodology is named MOBMAS, which stands for ???Methodology for Ontology-Based MASs???. A major improvement of MOBMAS over the existing agentoriented MAS development methodologies is its explicit and extensive support for ontology-based MAS development. Ontologies have been widely acknowledged for their significant benefits to interoperability, reusability, MAS development activities (such as system analysis and agent knowledge modelling) and MAS operation (such as agent communication and reasoning). Recognising these desirable ontology???s benefits, MOBMAS endeavours to identify and implement the various ways in which ontologies can be used in the MAS development process and integrated into the MAS model definitions. In so doing, MOBMAS has exploited ontologies to enhance its MAS development process and MAS development product with various strengths. These strengths include those ontology???s benefits listed above, and those additional benefits uncovered by MOBMAS, e.g. support for verification and validation, extendibility, maintainability and reliability. Compared to the numerous existing agent-oriented methodologies, MOBMAS is the first that explicitly and extensively investigates the diverse potential advantages of ontologies in MAS development, and which is able to implement these potential advantages via an ontology-based MAS development process and a set of ontology-based MAS model definitions. Another major contribution of MOBMAS to the field of AOSE is its ability to address all key concerns of MAS development in one methodological framework. The methodology provides support for a comprehensive list of methodological requirements, which are important to agent-oriented analysis and design, but which may not be wellsupported by the current methodologies. These methodological requirements were identified and validated by this research from three sources: the existing agent-oriented methodologies, the existing evaluation frameworks for agent-oriented methodologies and conventional system development methodologies, and a survey of practitioners and researchers in the field of AOSE. MOBMAS supports the identified methodological requirements by combining the strengths of the existing agent-oriented methodologies (i.e. by reusing and enhancing the various strong techniques and model definitions of the existing methodologies where appropriate), and by proposing new techniques and model definitions where necessary. The process of developing MOBMAS consisted of three sequential research activities. The first activity identified and validated a list of methodological requirements for an Agent Oriented Software Engineering methodology as mentioned above. The second research activity developed MOBMAS by specifying a development process, a set of techniques and a set of model definitions for supporting the identified methodological requirements. The final research activity evaluated and refined MOBMAS by collecting expert reviews on the methodology, using the methodology on an application and conducting a feature analysis of the methodology.

Thesis (MCIS - Computer and Information Sciences) -- AUT University, 2007. / Primary supervisor: Anne Philpott. Co-supervisor: Professor Stephen MacDonell. Includes bibliographical references. Also held in print (x, 147 leaves : ill. ; 30 cm.) in City Campus Theses Collection (T 005.12 LIC)

Diss. Göteborg : Univ., 2004.

Thesis(M.S.)--Auburn University, 2005. / Abstract. Vita. Includes bibliographic references (p.87-92).