Editorial

The sixth issue of ComSIS is in front of you. I am glad to say that we have gained a trust of the authors not only from the region, but all over the world. I believe that we will keep the continuity and strengthen its international character. Also, it is my great pleasure to announce that a new member joined ComSIS Consortium. Since September 2006, the School of Computing form the Union University of Belgrade has joined us, and Prof. Stevan Milinković, the Dean of the School, has become a member of the Managing Board. We hope that his support will significantly contribute to the ComSIS mission.

This issue of ComSIS contains two invited, and four regular research papers. The invited papers come from the distinguished authors, Miha Grčar, Dunja Mladenič, and Marko Grobelnik from the Jožef Štefan Institute from Slovenia, and Jozo Dujmović and Haishi Bai from the Department of Computer Science of San Francisco State University. The authors state that we live in a society in which computers and the Internet are widely used for accessing different kinds of information. Therefore, it is not enough that computers are capable of processing and storing huge amounts of data, but they also need to be accessible to a wider community. In this way, one of the important and even more interesting research areas in computer science is the development of natural, adaptive and intelligent interfaces and search engines for accessing data. Another one is the evaluation and selection of software systems, which is a complex problem characterized by a large number of heterogeneous inputs.

Miha Grčar, Dunja Mladenič, and Marko Grobelnik in their paper "User Profiling for the Web" address the problem of personalized information delivery related to the Web that is based on using profiling. They analyze three main groups of approaches to Web personalization: content-based filtering, collaborative filtering and Web usage mining. They provide an overview of them including recent research results in the area with the emphases on user profiling in the context of Semantic Web applications. One of the conclusions is that the user models are built automatically or semi-automatically using machine learning and data mining methods. One of the most important challenges for building user-models is an efficient semi-automatic mode, where only limited amount of human time is available for providing answers to different questions.

During the last ten years the search technology has made a dramatic progress and currently search engines are the most influential Web tools. Jozo Dujmović and Haishi Bai in their paper "Evaluation and Comparison of Search Engines Using the LSP Method" focus on building a comprehensive model for evaluation and comparison of search engines based on the general proprietary search technology (PST). The model reflects the capability of search engines to satisfy user requirements. The experimental part of the paper includes the evaluation of four leading PST search engines: Yahoo! Search, Ask, Google, and MSN. The authors stress that the evaluation of search engines is a complex software evaluation problem. It includes a spectrum of functionality, usability and performance inputs, including more than 80 individual quality attributes. They indicate the necessity of developing a specialized tool for measurement of response times, resource consumption, and the quality of information retrieval.

As a rule, we have at least one paper in each regular issue devoted to the research in AI. In the paper "A Comparison of the Bagging and the Boosting Methods Using the Decision Trees Classifiers", Kristína Machová, Miroslav Puszta, František Barčák, and Peter Bednár present an improvement of the precision of the classification algorithm results. The authors focus on the bagging and boosting methods in combination with the decision trees in the role of particular classifiers. This paper describes a set of experiments with bagging and boosting in the task of text categorisation. The results of performance tests are presented. The authors conclude that the bagging and boosting algorithms are suitable for increasing the efficiency of the classification algorithms, where boosting achieves better results than bagging, which is a simpler one.

Grey System theory is a new multidisciplinary theory dealing with systems for which we lack information. It is widely applied to solve the problems with uncertainty, where extracting and generating valuable information is needed. In the paper "Data Mining Techniques Based on Grey System Theories for Time Sequence Data", Liu Bin, Zhang Hui, Liu Sifeng and Dang Yaoguo propose data mining methods for time sequence based on the Grey system theories, and thoughts on data mining with embedded knowledge.

The next two papers represent the research area of security and protection, as an important discipline in computer science and IT industry. Constant increase in use of wireless networks created a need for strong safety mechanisms. Radomir Prodanović and Dejan Simić in their paper "Holistic Approach to Wep Protocol in Securing Wireless Network Infrastructure" argue that traditional mechanisms for physical protection of wired networks cannot be applied to protection of wireless networks. Therefore, it is a necessary to create new mechanisms for protection of wireless networks in order to enable using wireless networks in a safe way, with respect to accuracy and privacy. They describe the Wired Equivalent Privacy (WEP) protocol for protection of wireless networks, its security deficiencies, as well as various kinds of attacks that can jeopardize security goals. A summary of security improvements of WEP protocol is given, and the advantages of the new 802.11i standard are also discussed.

Bojan Jovičić and Dejan Simić, the authors of the paper "Common Web Application Attack Types and Security Using ASP.NET" focus on the attacks against Web applications, either to gain direct benefit by collecting private information, or to disable target sites. They discuss the two most common Web application attacks: SQL Injection and Cross Site Scripting, and propose how to use ASP.NET to provide Web applications security.

On behalf of the ComSIS Consortium, I would like to take this opportunity to give great thanks to the reviewers and all of the authors for their high-quality work, great efforts, and remarkable enthusiasm.

Editor-in-Chief
Ivan Luković