completeness in the modeling of contexts, but rather to provide easy-to-use means to define basic categories of contexts.
The model is based on the abstract notion of spheres. Generally, spheres can be described as a kind of classification of contexts which are agreed upon by the users and within which actions manipulating particular artifacts may occur (e.g. the financial administration of project A within department B). Spheres can be configured jointly and are ordered hierarchically. Co-workers use such collectively defined spheres to focus on specific aspects of their cooperative work. The configuration of a sphere is very flexible and can represent organizational contexts (units, departments, etc.) as well as social or physical ones (time, space, etc.). A sphere contains a set of jointly used artifacts, possible sub-spheres and a set of contextors which represent the most fine-grained level of a particular sphere. While spheres represent particular contexts, contextors represent actions (which may be related to tasks) within this specific context. In performing an action, or a set of actions, on an artifact which is represented in a particular sphere, the performing user can choose a contextor to classify the performed action to a fine-grained level (e.g. "fixing the annual report" or "auditing of accounts"). Regarding events contextors and their superior spheres can be understood as cues of events.
From this follows that contexts within Atmosphere are generally represented under two different aspects. Firstly the decision to locate an artifact in a specific sphere is an artifact-related contextualization that relates the artifact to a context defined jointly by the users. Secondly, performing specific actions on artifacts, while selecting pre-defined contextors, generates awareness information that is located on a higher level of aggregation than basic events on artifacts (like opening, deleting etc). It is thus possible to make users aware of cues of actions rather than single events and to represent part of the intentions of the performer.
The Atmosphere model is to support three major design goals: (1) to enable the configuration of contexts: although individual working contexts could still be useful, most of the jointly used working contexts are used by groups. Therefore, methods for defining and negotiating contexts have to be provided. (2) to enable the selection of contexts: assigning information about actions to working contexts involves a huge amount of extra work for the users. A pool of methods like annotations, contextors and default settings should help users to reduce extra work. (3) to enable the presentation of contexts: while working in different contexts, the relevant spheres have to be visible and easily manipulable. Awareness information assigned to spheres has to be displayed by selected user- defined criteria.
In order to realize these design goals, the Atmosphere model comprises three basic architectural elements. The Sphere Browser provides means to display and