XRCE explores the new world of pervasive information technology through several related research efforts, two of which specifically address pervasive IT: Devices and Intermediation and Contextual Computing. The Contextual Computing research area aims at making software applications context-aware. A context-aware application is one that adapts its behaviour to its users and their situation.

In a related research effort, Devices and Intermediation , we focus on the ubiquitous computing environment, where traditional computers are being replaced or complemented by a multitude of other devices, either wearable (phones, PDAs, intelligent wristwatches, smart clothing, etc.) or embedded in the environment (public displays, sensors, intelligent appliances in the home, office or car, etc.). While ubiquitous computing certainly offers completely new opportunities for user applications, there are many technical barriers to overcome, such as interoperability between devices, security, privacy and maintainability of applications.

About context

A definition of context as taken from a dictionary may state that "The context of an idea or event is the general situation that relates to it and which helps it to be understood." The study of a system vs. its context is an important aspect of systems theory, where a fundamental question is where to set the boundary of a system. The general approach is to say that the boundary delimits that part for which we seek a detailed understanding in terms of the constituent components, their behaviours and relationships. What is inside the boundary is "the system". The "context" or "environment" is everything that is outside the boundary, but that may interact with the system and influence its behaviour. It is often pointed out that where to set the boundary is a purely subjective decision. There is no intrinsic feature of an element that decides if it is part of a system or not. Thus the context of one system can be an integral part of another system with a different boundary.

Contextual computing

Contextual computing deals with how contextual information about users, documents, and systems can be taken into account to enhance application services. This should be compared to the situation today, where most applications often behave the same way regardless of who or where you are, who you're with and what you're currently doing. For example, a human observer would notice individual usage patterns, such as the fact that when I undock my laptop computer in the evening before leaving the office I always synchronize my calendar first, but not if I undock it before a meeting. Current applications would not be able to adapt their behaviour accordingly, without some fairly complex programming. The role of contextual computing is to extend current systems with abilities to detect contextual information, to represent it, to manipulate it (maybe replacing old patterns with new ones), and to influence the behaviour of applications in order to better support users. In the example just cited, the useful context of the undock operation is time (evening or not) and the content of the personal agenda of the user (meeting scheduled or not). Depending on the context, the undock operation would then be able to behave differently.

Contextual computing at XRCE

In this research program, we study how applications and functions can benefit from context and how they can technically be made context-aware. We take a fairly broad view on what aspects can contribute to the context. In some cases, context can be picked up via software probes in existing applications (e.g. to generate and analyse a Web access pattern), but we expect that context information will increasingly be made available from emerging ubiquitous computing environments. These environments rely on large numbers of interconnected devices, some of which are mobile with the users, others embedded in the physical environnment.

A common goal for our work is that rather than having contextual information to be explicitly input by the users, it is sensed and inferred by a software infrastructure, which also aims to respect the privacy of users. Devices not only give information access to users (via mobile phones and personal digital assistants), but can also provide position sensing, face and gesture recognition, etc. The contextual computing research is carried out in a multi-disciplinary team with background in computer science and sociology.

Applications that can benefit from context

Some commercial applications that take context into consideration are beginning to appear, especially in the domain of mobile phones. With the emergence of low-cost positioning based on GPS or triangulation, location-based services now provide information about nearby shops and restaurants. While current location-sensing, and context sensing in general, is technology dependent, future services will offer abstractions of context to applications, which will be able to use these services in a less technology dependent ways. In the professional environment we can expect a large variety of new context-aware services in the near future, for example:

• categorisation and ranking of documents, links, and other information
• information filtering
• presentation of information to users
• content generation, for example summarization or added explanations and references
• adaptation of presentation to media, for example adaptation to small screens, or adaptation for printing (eg. by translation of hyperlinks to footnotes or bibliographic references)

Projects

The reasearch is currently carried out within the following projects:

• Contact. The goal of this project is to build a platform for context awareness that allows different sensing mechanisms to be dynamically selected and combined, and to offer higher levels of abstractions of sensed data to users and applications.
• MILK. This is an EU funded project that develops context-aware applications for information sharing in distributed organisations. XRCE particularly focuses on the use of context-aware shared large screens in semi-public places. This project builds on results from the Campiello project.
• Web Information Discovery. This project explores the Hidden Web, i.e. the various information content that is available through the Web, but which is "hidden" behind query portals, which prevents the information from being available to the ordinary Web crawlers and indexing systems that are used by the normal Web serach engines.

For further contact about the project, please contact Christer Fernstrom