Zoltán Fiala

Engineering the Presentation Layer of Adaptable Web Information Systems

Engineering adaptable Web Information Systems (WIS) requires systematic design models and specification frameworks. A complete model-driven methodology like Hera distinguishes between the conceptual, navigational, and presentational aspects of WIS design and identifies different adaptation “hot-spots” in each design step. This paper concentrates on adaptation in the presentation layer and combines the modeling power of Hera with the versatile presentation capabilities of the AMACONT project. After discussing different aspects of presentation layer adaptation, the layout manager mechanism of AMACONT for describing the adaptable layout of ubiquitous Web presentations is introduced. Then the RDFS-based Hera schema for presentation models is presented, allowing to assign AMACONT layout descriptors to Hera slices. According to this formalization, Hera application model instances are automatically transformed to a component-based AMACONT implementation that can be adjusted to different end devices and output formats. The XML-based transformation process is explained in detail, and the resulting methodology is exemplified by a prototype application.

Personalization-Based Optimization of Web Interfaces for Mobile Devices

Developing personalized applications for the ubiquitous Web assumes to provide different user interfaces addressing heterogeneous capabilities of device classes. Major problems are the lack of sufficient presentation space and the diversity of interaction techniques, both requiring adaptive intelligent user interfaces. To meet this challenge this paper introduces an approach for the personalization-based optimization of Web interfaces for mobile devices. On the basis of a user model different adaptation issues are discussed. Firstly, static adaptation mechanisms affecting the structure of Web documents as well as layout managers enabling a device independent definition of Web presentations for heterogeneous devices are introduced. Then an interactive mechanism for dynamically predicting user preferences for hiding unnecessary information through content adaptation is presented. As a proof of concept an architecture realized by a pipeline-based document generator was developed for static/dynamic adaptation, which is partly explained in this paper.

Adaptation and reuse in designing Web information systems

The increasing number of requirements for a Web information system asks for an engineered process in designing such a system. In this paper we focus on two of these requirements: presentation adaptation based on user preferences/device capabilities and reusability of the different design artifacts. Hera is a model-based design methodology for Web information systems. Adaptation and reuse can be tackled at different design levels in Hera. We illustrate by means of examples how adaptation and reuse can be achieved in the conceptual model, the application model, and the presentation model. Based on a static user profile, adaptation is realized by attaching appearance conditions to model elements. With respect to reusability we do focus on one mechanism that supports it, namely inheritance.

Design and implementation of component-based adaptive Web presentations

Engineering adaptive Web applications implies the development of content that can be automatically adjusted to varying client devices and user preferences. To meet this requirement, the AMACONT project recently introduced a component-based XML document format. Configurable document components encapsulating adaptive behavior and layout are used on different abstraction levels in order to support flexible reuse for effective Web page generation. This paper focuses on the process of designing and implementing such component-based adaptive Web presentations. Based on the model-driven specification framework from the Hera project, different phases of adaptation design are identified and their realization using AMACONT components is explained. Finally, a pipeline-based document generator for dynamically publishing component structures to different Web output formats is described.

Considering additional adaptation concerns in the design of web applications

The design of Web applications traditionally relies heavily on the navigation design. The Web as it evolves now brings additional design concerns, such as omni-presence, device-dependence, privacy, accessibility, localization etc. Many of these additional concerns are occurrences of user- or context-dependency, and are typically realized by transformations of the application (design) that embed adaptation in the navigation. In this paper we focus on how to extend an application with new functionality without having to redesign the entire application. If we can easily add functionality, we can separate additional design concerns and describe them independently. Using a component-based implementation we show how to extend a Web application to support additional design concerns at presentation generation level. Furthermore, we demonstrate how an Aspect-Oriented approach can support the high-level specification of these (additional) design concerns at a conceptual level.

Developing component-based adaptive Web applications with the AMACONTBuilder

The growing need for personalization and device independence calls for effective ways of engineering adaptive Web applications. This requires formats, languages and structured process models that allow developers to design and express adaptation, but also appropriate authoring tools. Still, current Web authoring tools do not provide sufficient adaptation support. Recently, the AMACONT project introduced a component-based document format aiming at efficiently composing adaptive Web applications from reusable Web components. This paper focuses on the development of such applications and presents a visual authoring tool called AMACONTBuilder. Based on a structured authoring process it is shown how it can be utilized for effectively creating AMACONT applications.

From adaptation engineering to aspect-oriented context-dependency

The evolution of the Web requires to consider an increasing number of context-dependency issues. Therefore, in our research we focus on how to extend a Web application with additional adaptation concerns without having to redesign the entire application. Based on a generic transcoding tool we illustrate here how we can add adaptation functionality to an existing Web application. Furthermore, we consider how an aspect-oriented approach can support the high-level specification of such additional concerns in the design of the Web application.