Technique for dynamic composition of content and context-sensitive mobile applications. Adaptive mobile browsers as a case study: Dissertation

The mobile environment brings new challenges for applications. Mobile usage is spontaneous and applications should be fast to install, start, and use in mobile devices and wireless networks. The wireless network connections offer typically less bandwidth than fixed line connections and may cause costs for the user. In addition, the input and output capabilities and memory and processing power resources of mobile devices are typically limited in comparison to desktop computers. This all sets requirements for adaptation methods that could provide more usable and efficient applications for specific users, contexts, and services available on the Web. Implementation of adaptive applications requires methods for context-sensing and adaptation. The context can change rapidly when a user is moving in a physical environment. Hence, methods that can fast adapt an application for a rapidly changing context are needed. An adaptive application should learn about user behaviour, sense the activity of the user, and use the idle time of the application for speculative adaptation that prepares application parts for potential future contexts in the background. In addition, errors may arise while an adaptive application is being composed for a new context. For example, if a mobile device is disconnected, it is not possible to fetch contents from the Web. In this dissertation it is argued that the task-based composition technique helps developers to construct adaptive applications for mobile devices and makes the dynamic composition of content and context-sensitive applications more fluent. The task-based composition technique is based on the model of the World Wide Web Consortium (W3C) that provides a requestor-adaptor structure for content adaptation. Like the requestor-adaptor element of the W3C model, a task is an adaptation element that can provide additional context information, request other tasks, adapt their responses, and deliver new or refreshed responses for the requestors. Tasks can prepare content and context-sensitive application instances for current and predicted contexts in many phases and finally compose an application of the prepared instances. The task-based composition technique extends the W3C model with task factories that construct tasks for adaptation requests and specific contexts. Tasks are defined with an XML-based language that enables developers to describe tasks and context-sensitive adaptation actions and their settings. Both context-sensitive tasks and application instances can be cached, which can speed up the adaptation of applications. This dissertation focuses on adaptive browsers that are constructed for mobile devices and discusses how the task-based composition technique can support client-side dynamic composition of content and context-sensitive applications and improve performance when UIs are adapted for rapidly changing contexts and for services available on the Web.