Where's Waldo? - or - A taxonomy for thinking about location in pervasive computing

Abstract

[Introduction] Virtually all pervasive computing systems use some form of location for affecting the system's behaviour. Location-based services are available commercially, albeit in a primitive form, from many mobile telecommunications providers. Despite this, location is a remarkably subtle concept to reason with. There are a huge number of possible answers to what is superficially a simple question: each kind of answer reveals something about the way in which we conceptualise location and any services based upon it. This may seem like a trivial observation, but it cuts to the heart of the development of programming environments for pervasive computing. Software design is the process of mapping concepts to programming structures, whether one is developing object models or complete programming languages. A system is easy to program when the developer can express his concepts directly or with minimal mental gymnastics; conversely, the developer takes on a considerable load when using a system without the requisite concepts. Handling location is a basic requirement for many pervasive computing systems: but what kind of location will we provide in a programming environment? If we choose the "wrong" representation we will end up with an awkward (or unusable) environment. In this paper we try to exhaust the ways one may answer a simple location question in a pervasive computing system, to develop a taxonomy of answers that an application might want to receive. We deliberately stray beyond the bounds of traditional location systems and the information typically at hand in order to see how a "really pervasive" computing system might view location. We analyse the space of answers to see what conclusions can be drawn for the development of semantic models and programming environments.