No abstract

This paper presents Geo-Linda, a physically distributed tuple space. Geo-Linda targets ubiquitous computing applications involving the detection of movement patterns of objects and people, such as the meeting of two people or the insertion of a product in a shopping cart, the loading of container in a boat. . . Existing distributed tuple spaces suffer two main limitations: (1) they cannot detect precisely movement patterns, leading to inconsistencies between the state of applications and the physical world; (2) they cannot detect several kinds of movement patterns. Geo-Linda tackles these two limitations, by introducing a geometric addressing mode and two new reading operations to access the tuple space. In this paper, we illustrate Geo-Linda with several small examples and a full application.

In this article we present an application to navigate through a photo collection, using context. This application proposes two navigation schemes: physical and virtual navigation. Physical navigation permits the user to seamlessly access to photos that were taken near him. Virtual navigation permits the user to virtually explore the surroundings by jumping from one photo to another.This application relies on a definition of context which is based on a notion of proximity. Proximity can be determined according to one or several dimensions. For example, with this definition the user's context can be represented by the photos that were taken near him during a precise period like the Second World War. Such a context permits the user to physically navigate the photo collection nowadays and to get photos of the surroundings that were taken during another time period. Our application offers other services, such as temporal navigation to get different views of a place, route planning or virtually entering into closed buildings.