Vol.8 No.1 March 1, 2009       
 Design of Sophisticated Web-based Systems

Editorial (pp001-002)
        L. Baresi, P. Fraternali, and G.-J. Houben

Research articles:
Enriching Information Retrieval Results with Web Accessibility Measurement (pp003-024)
        Markel Vigo, Myriam Arrue, and Julio Abascal
Search engines are the most common gateway to search information in the WWW. Since Information Retrieval (IR) systems do not take Web accessibility issues into account, displayed results might not tailor to certain users’ needs such as people with disabilities or mobile devices’ users. In order to overcome this situation, we present a model aiming at considering Web accessibility as well as content relevance. The model consists of three components (Content Analysis Module, Accessibility Analysis Module and Results Collector Module) that carry out the following tasks: content analysis, automatic Web accessibility evaluation and accessibility measurement of results for re-ranking. Since criteria for ranking results provided by IR systems are necessary, quantitative metrics for accessibility have also been defined. Two prototypes that follow the specifications of the model have been developed in order to demonstrate the feasibility of this proposal. Finally, some case studies have been conducted aiming at discovering how traditional search engines deal with Web accessibility.

Oblivious Integration of Volatile Functionality in Web Application Interfaces (pp025-047)
        Jeronimo Ginzburg, Daniano Distante, Gustavo Rossi, and Matias Urbieta
Web applications are used to fast and continuous evolution. In response to new or changing requirements, additional code is developed and existing one is properly modified. When new requirements are temporary, i.e., when they specify some volatile functionality that is expected to be online only for some time and then removed, the additions and changes are destined to be later rolled back. This way to proceed, apart from being time and effort demanding, by involving the intrusive editing of the application’s source code, brings along the risk of polluting it and introducing mistakes. In this paper, we present an approach to deal with volatile functionality in Web applications at the presentation level, based on oblivious composition of Web user interfaces. Our approach, which is inspired by well-known techniques for advanced separation of concerns such as aspect-oriented software design, allows to clearly separate the design of the application’s core user interface from the one corresponding to more volatile functionality. Both core and volatile user interfaces are oblivious from each other and can be seamlessly composed using a transformation language. We show that in this way we simplify the application’s evolution by preventing intrusive edition of the user interface code. Using some illustrative examples, we focus both on design and implementation issues, presenting an extension of the OOHDM design model that supports modular design of volatile functionality.

Fine-Grained Specification and Control of Data Flow in Web-Based User Interfaces (pp048-070)
        Matthias Book and Volker Gruhn
When building process-intensive web applications, developers typically spend considerable effort on the exchange of specific data entities between specific web pages and operations under specific conditions, as called for by business requirements. Since the WWW infrastructure provides only very coarse data exchange mechanisms, we introduce a notation for the design of fine-grained conditional data flows between user interface components. These specifications can be interpreted by a data flow controller that automatically provides the data entities to the specified receivers at run-time, relieving developers of the need to implement user interface data flows manually.

Finding Unexpected Navigation Behaviour in Clickstream Data for Website Design Improvement (pp071-092)
        I-Hsien Ting, Chris Kimble, and Daniel Kudenko
This paper describes a novel web usage mining approach to discover patterns in the navigation of websites known as Unexpected Navigation Behaviours (UNBs).  The approach provides a web designer with a means of identifying and classifying patterns of browsing and, by reviewing these patterns, the designer can then choose to modify the design of their site or redesign it completely.  UNB mining is based on the Consecutive Common Subsequence (CCS), a special instance of Common Subsequence (CS), which is used to define a set of expected routes.  The predefined expected routes are then treated as rules and stored in a rule base.  By using the predefined route and the UNB mining algorithm, interesting navigation behaviours can be discovered.  This paper will introduce the format of the expected route and describe the UNB algorithms.  It will also describe a tool that a website designer can use to define the expected route more efficiently, which can help the website designer to make decision about where and how the design of website can be improved.  The paper concludes with a series of experiments designed to evaluate how well the UNB mining algorithms work and demonstrate how UNB mining can be useful for improving website design.

Back to JWE Online Front Page