Salesian Journal on Information Systems

3

Authors: Finger, M.

We show that recent known advances in Computer Science have allowed us to enter in regions of data "imensity" that we thought were unreachable. Besides, formal and precise models of specific parts of the worls that we are building are increasingly detailed and allow us to simulate more accurately both human and natural phenomena. Paradoxically, we argue that this growth both of size and precision will take us to a Computer Science that is closer to humanization.

6

Authors: BARBOSA, G. C.; SILVA, B. Z.; DALPRA, H. L. O; VILARINO, I. F., PAIVA. M. M., ARBEX, W.

Distributed computing systems are used for high performance computing tasks, taking advantage of join processing power of multiple independent computers, but they are interconnected by network. These high performance systems can be divided into two classes of distributed computing systems: computer cluster and grid computing. The FightAIDS@Home is a distributed system toward the search for solutions for the AIDS treatment, and this project is led by the Olson Laboratory, California. Nowadays, this project use computing resources of the Grid World Community Grid, which consist mostly of computers of ordinary users, using open and non-specific standard protocols and interfaces to ensure interoperability between different systems. These paper presents informations from a case study about FightAIDS@Home and its implementation, using aspects of distributed systems such as grid computer and cloud computing to support implementation of collaborative computing.

10

Authors: YOKOME, E. A.; ARANTES, F. L.

Ontologies have been used in various research areas from computer science, including data mining. This article describes the development of a domain ontology for data mining. Meta-DM ontology provides a common terminology that can be shared and processed by data mining tools. The ontology also indicates the KDD phases where human knowledge is necessary - this is its greatest differential. With ontologies as the Meta-DM, it is possible to add semantics in the data mining process in order to improve the interaction and cooperation between experts and mining systems.

4

Authors: VOLPATO, G. L.

In this work, I describe a logical method for scientific writing. Any decision made in this type of writing should be based on the logic of science and the rules of communication, as part of a creative discourse. I present some logical flaws (regarding journal classification, academic vs. non-academic texts, and subjective inferences) and writing mistakes (in the structure of a paper and the writing style) that can undermine publication.

10

Authors: AFFONSO, F. J., RODRIGUES, E. L. L.

The software development process has been directed, over the years, to various methodologies with specific purposes to attend emerging needs. Besides, it can also be noticed, during this period, that some processes require mechanisms related to software reuse and greater speed in the development stage. An important factor in this context is the mutation (adaptation), which occurs in all the software's life cycle, due to its customers' needs or due to technological changes. Regarding the latter factor, it has been observed a significant increase in developments that use distributed applications through the World Wide Web or remote application. Based on the adaptation idea and on the necessity of software distribution systems, this paper presents a technique to reconfigure software capable of acting in several developmental contexts (local, distributed and/or Web). In order to demonstrate its applicability, a case study, through the use of service orientation and remote calls, was done to show the software adaptation in the development of applications. Besides, comparative results among the approaches used in the development of reconfigurable applications are also presented.

15

Authors: REZENDE, S. O.; MARCACINI, R. M.; MOURA, M. F.

The progress in digitally generated data aquisition and storage has allowed for a huge growth in information generated in organizations. Around 80% ofthose data are created in non structured format and a significant part of those are texts. Intelligent organization of those textual collection is a matter of interest for most organizations, for it speed up information search and retrieval. In this context, Text Mining can transform this great amount non structure text data un useful knowledge, that can even be innovative for those organizations. Using unsupervised methods for knowledge extraction and organization has received great attention in literature, because it does not require previous knowledge on the textual collections that are going to be explored. In this article we describe the main techniques and algorithms used for unsupervised knowledege extraction and organization from textual data. The most relevant works in literature are presented and discussed in each phase of the Text Mining process and some existing computational tools are suggested for each task at hand. At last, some examples and applications are present to show the use of Text Mining on real problems.