The software development life-cycle is comprised of a series of successive activities consisting of analysis, design, implementation, system testing and maintenance. During the analysis phase we do planning and requirements definition for the software product. The design phase, which follows the analysis phase, is concerned with deciding exactly how the software will be implemented. However, the actual transition from the analysis to the design phase is not well documented. There exists an information gap between these two phases.In this study, the transition from the analysis to the design phase is explored by using the reverse engineering method which essentially proceeds from the design phase back to the analysis phase. This study is based on the design of an approximately five thousand line project - an Executive Calendar, which is first designed using a computer-aided software engineering (CASE) tool called DesignAid. The transition is documented in order to exploit the isomorphisms between each phase.The end results show that by documenting the mapping between the analysis phase and the design phase, the process of transition from one phase to another could be partly automated. By using the reverse engineering method, the elements which are necessary in the transition between the analysis and the design phase can be easily identified. Being able to identify these elements, one can reduce the amount of effort required to transform user requirements to design, and thus improve software productivity. / Department of Computer Science

A major reengineering goal is software reuse. Effective reuse of knowledge, processes and products from previous software developments can reduce costs and increase both productivity and quality in software projects.This thesis extensively tests five projects produced by the graduate software engineering class at Ball State University. Each project has the same set of requirements.Each project is also analyzed based on subjective criteria, for example documentation, use of mnemonics for variable names and ease of understanding. Based on the outcome of testing and subjective analysis, reusable parts are identified.Metrics are collected on all of these projects. This thesis compares the metrics collected on the modules identified for reuse, and the same metrics collected on the non-reusable modules, to determine if there is a statistically significant difference in those metrics between the two groups. Metrics which are good predictors of reusable modules are identified.Metrics which are found to be good predictors of reusable modules include: number of in-parameters, number of data structure manipulations and central calls. / Department of Computer Science

This thesis proposes a novel approach, called Analogy-X to extend and improve the classical data-intensive analogy approach for software effort estimation. The Analogy-X approach combines the notions of distance matrix correlation found in ecology literature and statistic analysis techniques to provide useful inferential statistics to support analogy-based systems. Data-intensive analogy for software effort estimation has been proposed as a viable alternative to other prediction methods such as linear regression. In many cases, researchers found analogy outperformed algorithmic methods. However, the overall performance of analogy depends on the dataset quality or relevance of project cases to the target project, and the feature subset selected in the analogy-based model. Unfortunately, there is no mechanism to assess its appropriateness for a specific dataset, in most of the cases analogy will continue to execute regardless of the dataset quality. The Analogy-X approach is a set of procedures that utilize the principles of Mantel randomization test to provide inferential statistics to Analogy. Inspired by the Mantel correlation randomization test commonly used in ecology and psychology, Analogy-X uses the strength of correlation between the distance matrix of project features and the distance matrix of known effort values of the dataset to assess the suitability of the dataset for analogy, to identify the most appropriate feature subset, and to remove any atypical project cases from the dataset. The empirical studies show that Analogy-X is capable of: -- Detect extremely outlying project cases that will ultimately distort prediction outcomes using a sensitivity analysis strategy. -- Detect relevant project features that are useful to identify potential source analogues in a stepwise fashion similar to that of stepwise regression. -- Identifying whether analogy-based approach is appropriate for the dataset Analogy-X, thus is a robust solution, provides a sound statistical basis for analogy. It removes the need of using any forms of heuristic search and greatly improves its algorithmic performance. The studies also show that the Analogy-X approach is capable of removing the bottlenecks of performance in data-intensive analogy. The overall results obtained also suggest that a fully automated data-intensive analogy for software effort estimation can be implemented using the Analogy-X approach, and it is indeed an effective front end to analogy-based systems. The contribution of this work is significant since it provides an approach that will have major impact on the evolution of data-intensive analogy-based and case-based reasoning systems.

Thesis (Ph. D.)--University of Washington, 1997. / Vita. Includes bibliographical references (leaves [144]-153).

Thesis (M.S.)--West Virginia University, 1999. / Title from document title page. Document formatted into pages; contains xi, 125 p. : ill. (some col.). Includes abstract. Includes bibliographical references (p. 108-111).

Thesis (M.S.)--Oregon State University, 2010. / Printout. Includes bibliographical references (leaves 80-87). Also available on the World Wide Web.

Made available in DSpace on 2013-08-07T18:42:39Z (GMT). No. of bitstreams: 1 000448243-Texto+Completo-0.pdf: 2290128 bytes, checksum: 5d6aa8c3c9b323bf5e5fc7c54792fa85 (MD5) Previous issue date: 2010 / Nowadays, a considerable part of usability evaluations carried out during the software development lifecycle is conducted at later stages. Some of these assessments depend on the creation of fully functional prototypes, so they can be applied, and they’re commonly performed after the development and deployment of software. This factor contributes significantly to the increase in cost of production. In this sense, from the research of prototyping, usability engineering and the conduction of a systematic review, this work presents an approach that allows the usability evaluation anticipation, through the use of mixed-fidelity prototypes created via the use of a computational tool. This tool assists in the creation and simulation of prototype interfaces, and enables collecting the following metrics of usability, for further analysis: task success, giving up on tasks, time on task, effort, screen balance, screen symmetry, screen balance, sequencing, order and complexity and system usability scale. In addition to the metrics that are collected by the tool automatically, other metrics are proposed to be collected from the observation of users using the prototypes created: partial success of the task, number of errors, help and documentation, time help, anger or frustration. The analysis of the obtained metrics point to areas of the system with potential usability problems. Moreover, the execution of new tests using the interfaces refined with the approach aid allow us to quantify the improvements resulting of the layout changes. In addition to the approach, the work presents its use in an experiment conducted with users. This experiment showed that, although not ali the metrics that can be collected from prototypes contribute to the refinement of interfaces, it was still possible to reduce the number of usability problems by just analyzing the results of the metrics. / Hoje em dia, uma considerável parte das avaliações de usabilidade realizadas nos ciclos de desenvolvimento de software é conduzida em etapas mais avançadas nestes ciclos. Algumas destas avaliações dependem ainda que protótipos totalmente funcionais sejam construídos, para que possam ser aplicadas, sendo comumente realizadas após o desenvolvimento e implantação do software. Este fator contribui significativamente para o aumento no custo de produção. Neste sentido, a partir de pesquisas sobre prototipação e engenharia de usabilidade e da condução de uma revisão sistemática sobre métricas para avaliação de usabilidade, o presente trabalho apresenta uma abordagem que permite que a avaliação de usabilidade seja antecipada, efetuando-a por meio da utilização de protótipos — de fidelidade mista - elaborados mediante o uso de uma ferramenta computacional. Esta ferramenta auxilia na criação e simulação de protótipos de interfaces, além de permitir coletar as seguintes métricas de usabilidade, para posterior análise: sucesso de tarefas, desistência de tarefas, tempo na tarefa, esforço, equilíbrio de tela, simetria de tela, balanço de tela, seqüenciamento, ordem e complexidade e escala de usabilidade do sistema. Além das métricas que são coletadas pela ferramenta, de forma automática, são propostas as seguintes métricas, a serem coletadas a partir da observação dos usuários utilizando os protótipos criados: sucesso parcial de tarefa, número de erros, auxilio e documentação, tempo de ajuda e raiva ou frustração. A análise das métricas obtidas permite apontar áreas do sistema com possíveis problemas de usabilidade. Além disto, a execução de novos testes utilizando as interfaces refinadas com o auxílio da abordagem permite quantificar as melhorias realizadas pelas mudanças de layout efetuadas. Além da abordagem, o trabalho apresenta seu uso em um experimento realizado com usuários. Este experimento permitiu verificar que, apesar de nem todas as métricas passíveis de serem coletadas de protótipos contribuírem para o refino das interfaces, ainda assim foi possível reduzir o número de problemas de usabilidade.

Made available in DSpace on 2013-08-07T18:42:45Z (GMT). No. of bitstreams: 1 000418602-Texto+Completo-0.pdf: 874285 bytes, checksum: 2f440931163dba272efc1b7262dc5cf5 (MD5) Previous issue date: 2006 / The speed up growth and the market competition have taken companies to search for different ways to continuously increase quality and productivity in its software development processes. One way to optimize a software development process is the adoption of software reuse techniques, such as Component Based Development (CBD), which concerns productivity increase and cost reduction with quality. But it will be the component based development, by itself, a guarantee of economic benefit, or either, the costs to buy or integrate those components will always be lesser than the development cost of the same functionalities for one given application? Will the development of reusable components be a guarantee for recovering the accomplished investment from its creation? In this context, this work presents a proposal for a viability analysis model of reusable components creation, based on a series of qualitative and quantitative inputs, which can provide investment return projections for the creation of reusable components. Such model intends to serve as an assistance tool to the software engineers decision making process involved in reusable components development. / O acelerado crescimento e a competição pelo mercado têm levado empresas a buscarem formas de continuamente aumentar a qualidade e produtividade nos seus processos de desenvolvimento de software. Uma forma para se otimizar o processo de desenvolvimento de software é a adoção de técnicas de reuso de software como Desenvolvimento Baseado em Componentes (DBC) que abrange preocupações como o aumento da produtividade e a redução dos custos com qualidade. Mas será o desenvolvimento baseado em componentes, por si só, garantia de beneficio econômico, ou seja, os custos de compra/integração de componentes reutilizáveis serão sempre menores do que o custo de desenvolvimento das mesmas funcionalidades para uma dada aplicação? E o desenvolvimento de um componente reutilizável será garantia de recuperação do investimento realizado para criação do mesmo? Neste contexto, o presente trabalho apresenta uma proposta de modelo de análise de viabilidade para a criação de componentes, que com base em uma série de entradas qualitativas e quantitativas, provê projeções de retorno sobre o investimento na criação de componentes. Tal modelo tem como propósito servir como ferramenta de auxílio à tomada de decisão para os engenheiros de software envolvidos em programas de desenvolvimento de componentes reutilizáveis.

Made available in DSpace on 2013-08-07T18:42:48Z (GMT). No. of bitstreams: 1 000422501-Texto+Completo-0.pdf: 4374080 bytes, checksum: f654389731029dc20aa2d61a4b079c1d (MD5) Previous issue date: 2009 / Currently the main problem of Online Help Systems is that they tend to be directed to the specific needs and problems of their users. In the case of educational software, in particular, which deals with things related to teaching and learning, and its users include both students and teachers, with different goals and needs in relation to the same software, this problem is even greater. In this sense, the present dissertation, by researching about Online Help Systems in general, analyzing existing Online Help Systems for educational software, and by interviewing and testing its potential users, provides a way to structure and present the information contained in these systems, focusing ob what should be presented, in accordance with the features of this kind of software and its different profiles of users. / Atualmente, o principal problema de sistemas de ajuda on-line é que eles costumam ser pouco direcionados às necessidades e aos problemas específicos do usuário. No caso de software educacional, em particular, que trabalha questões relacionadas ao ensino e à aprendizagem, e seus usuários incluem tanto alunos quanto professores, tendo objetivos e necessidades diferentes em relação ao mesmo software, este problema é ainda maior. Neste sentido, o presente trabalho, a partir de pesquisas sobre sistemas de ajuda on-line em geral, da análise de sistemas de ajuda existentes para software educacionais e de entrevistas e testes com seus potenciais usuários, apresenta uma forma de estruturar e apresentar as informações contidas nestes sistemas, focalizando o que deve ser apresentado, de acordo com as especificidades deste tipo de software e de seus diferentes tipos de usuários.

Made available in DSpace on 2013-08-07T18:43:20Z (GMT). No. of bitstreams: 1 000424178-Texto+Completo-0.pdf: 2045392 bytes, checksum: 9acc5658b98379a3bd5e14c747a952da (MD5) Previous issue date: 2010 / The software industry invests amounts of money searching for the improvements on its products and services. Nowadays, clients are more demanding and look for industries that possess a severe quality control. That quality is measured through quality models adopted, such as CMMI, ISO, IEC, among others. The client’s demanding makes the organizations contract patterns that evaluate the quality level of its services. These contracts predict as punishments expensive fees. Observing of costs and deadlines is one of the requirements predicted in such contracts. The basis to its fulfillment (costs and deadlines), is the software development effort estimation and, after the delivery of the product, is the software maintenance effort. The challenge for researchers lies on supporting the software industry with effective works on estimation areas that may become much more accurate. Besides, the estimations shall not be economically heavy to their teams in order not to add more costs to the projects. This thesis comes from a research that has been developed through the last four years, from wich two of them were based on the actions in a software’s maintenance project from a great Brazilian public bank. The project was supported by a software industry partner of the Post graduation program in computer science from PUCRS. Such program inspired this thesis that presents as an option to the estimation areas a model of software’s maintenance effort. The model here presented supports the software industry as well the researches in the estimation area. / A indústria de software investe montantes de dinheiro em busca da melhoria de seus produtos e serviços. Atualmente, os clientes estão cada vez mais exigentes e procuram fábricas que possuam um rigoroso controle de qualidade. A qualidade é medida por meio dos modelos de qualidades adotados, como CMMI, ISO, IEC, dentre outros. A exigência dos clientes faz as organizações selarem contratos que medem o nível da qualidade de seus serviços. Esses contratos prevêem como sanções pesadas multas. O cumprimento de prazos e custos é uma das exigências previstas nesse tipo de contrato. A base para que se cumpram prazos e custos é a estimativa de esforço de desenvolvimento e, após a entrega do produto, o esforço para a manutenção do software. O desafio dos pesquisadores está em apoiar a indústria de software com trabalhos na área de estimativas que possam tornar as mesmas mais precisas. Além disso, as estimativas não devem onerar as equipes de software para não agregar mais custos aos projetos. Esta tese é fruto de uma pesquisa realizada ao longo de quatro anos, dois dos quais atuando em um projeto de manutenção de software pertencente a um grande banco governamental brasileiro. O projeto era mantido por uma fábrica de software parceira do Programa de Pós-Graduação em Ciência da Computação da Faculdade de Informática da PUCRS. A pesquisa inspirou esta tese de doutorado que apresenta, como uma opção para a área de estimativas, um modelo de esforço em manutenção de software. O modelo aqui apresentado apoia tanto a indústria de software como as pesquisas na área de estimativas.