Catálogo de publicaciones - libros

Semantic quality expresses the degree of correspondence between the information conveyed by a model and the domain that is modelled. As an early quality indicator of the system that implements the model, semantic quality must be evaluated before proceeding to implementation. Current evaluation approaches are based on ontological or meta-model analysis and/or use objective metrics. They ignore the model user’s perception of semantic quality, which also determines whether the benefits of using a faithful model will be achieved. The paper presents the development of a perceived semantic quality measure. It presents a measure pre-test, i.e. a study aimed at refining and validating a new measure before its use in research and practice. The results of the pre-test show that our measure is reliable and that it is sufficiently differentiated from other perception-based measures of information model use like ease of use, usefulness, and user information satisfaction.

As more and more companies are augmenting their data to include semantics, it is imperative that the choices made when choosing the modelling language are well founded in knowledge about the language and the domain in question. This work extends the Semiotic Quality Framework with computational and situated instruments. Furthermore, it demonstrates how the extended Semiotic Quality Framework can facilitate the choice of the most suited language for a real world application. The application is a directory services system, which currently is being moved into the realms of the Semantic Web.

In this paper a novel approach for the tool–based quality assurance of models is presented. The approach provides a meta model framework for domain specific and tool–independent quality assessment in heterogeneous model landscapes. In our framework we provide the concepts of queries, checks and views defined on meta model level and interpreted over the whole model landscape. Queries, checks and views are described in a predicative language based on the structures of the meta model.

Completeness is one of the important measures for semantic quality of a conceptual model, an ER model in our case. In this paper, a complete methodology is presented to measure completeness quantitatively. This methodology identifies existence of functional dependencies in the given conceptual model and transforms it into a multi-graph using the transformation rules proposed in this paper. This conversion can be helpful in implementing and automating computation of quality metrics for a given conceptual model. The new Fuzzy Completeness Index (FCI) introduced in this paper adopts an improved approach over Completeness Index proposed by authors in the previous research. FCI takes into account the extent a functional dependency has its representation in the conceptual model even when it is not fully represented. This partial representation of a functional dependency is measured using the fuzzy membership values and fuzzy hedges. The value of FCI varies between 0 and 1, where 1 represents a model that incorporates all the functional dependencies associated with it. Computation of FCI is demonstrated for a number of conceptual models. It is illustrated that the quality in terms of completeness can effectively be measured and compared through the FCI based approach.

We describe a new approach to managing information quality (IQ) in an e-Science context, by allowing scientists to define the quality characteristics that are of importance in their particular domain. These preferences are specified and classified in relation to a formal IQ ontology, intended to support the discovery and reuse of scientists’ quality descriptors and metrics. In this paper, we present a motivating scenario from the biological sub-domain of proteomics, and use it to illustrate how the generic quality model we have developed can be expanded incrementally without making unreasonable demands on the domain expert who maintains it.

Automated support for the requirements engineering (RE) process is a recognised research area. However, the practice still relies on office tools rather than RE-tools. Reasons include financial causes and difficulty to evaluate the available RE-tools. This work reports on an experiment trying to validate a previously proposed framework for evaluating RE-tools. The experiment participants used several alternative tools for making requirements specifications, and then evaluated the tools by means of the framework. This enables us to look at the participants’ performance with the various tools, evaluation approaches, and their perceptions about the same tools. The findings indicate advantages of using the evaluation framework, and of combining several evaluation techniques. The experiment indicates that RE-tools provide better support than office tools, leading to higher quality specifications.

Designers have accumulated much knowledge referring to OO systems design and construction, but this large body of knowledge is neither organized nor unified yet. In order to improve OO micro architectures, using the accumulated knowledge in a more systematic and effective way, we have defined a rules catalog (that unifies knowledge such as heuristics, principles, bad smells, etc.), the relationships between rules and patterns and an improvement method based on these subjects. We have carried out a controlled experiment which shows us that the usage of a rules catalog and its relationship with patterns really improves OO micro architectures.

Technology has changed the way we live, think and work. Technology has revolutionized business and now it must revolutionize teaching and learning. Live classroom based training is becoming too costly and cumbersome. 80 % of teachers and students already use computers. So e-learning can be seen as a means supporting life long learning with a lot of benefits. The worldwide web, high-capacity networks and high speed computers make learning available to people 24 hours a day in their office, at home or also in hotel rooms during a business trip around the globe. E-learning enables the access to learning when it is convenient.

The tutorial will provide an overview of FOOM methodology, which integrates the functional and object-oriented approaches. It will start with a rationale and need for the integrated approach, in light of the traditional methodologies on one hand, and the object-oriented and UML-based methodologies on the other hand. The main part of the tutorial will be devoted to a description of the analysis and design stages and sub-stages of FOOM, and their products. The description will be accompanied by a running example.

As the variability of information and software systems has increased, the need for an engineering discipline concerned with building reusable assets (such as specification sets, patterns and components) on one hand and representing and managing knowledge in specific domains on the other hand has become crucial. This discipline, called domain engineering, supports the notion of a domain, defined as a set of applications that use a set of common concepts for describing requirements, problems and capabilities. The purpose of domain engineering is to identify, model, construct, catalog, and disseminate a set of software artifacts that can be applied to existing and future software in a particular application domain. As such, it can support the effective and efficient management and development of software assets. Hence, it is important to introduce this discipline among software engineering practitioners and researchers.