The Cartesian approach to context

  • Authors:

  • John Plaice;Blanca Mancilla

  • Affiliations:

  • The University of New South Wales, Australia;The University of New South Wales, Australia

  • Venue:
  • Proceedings of the 2nd International Workshop on Context-Oriented Programming
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present a new approach for context-oriented programming in which the context is represented by a set of (dimension, value) pairs. This tuple parameterizes the environment, and it can be referred to either as a single entity or as a composed entity, parts of which can independently be accessed. The context is also an index into any programmable entity, in our model the hyperdatons, which are in turn, arbitrary-dimensional arrays of arbitrary extent. The context may have privileged dimensions and one such dimension is time, which has as well a physical interpretation. The importance of this dimension relies on the fact that its proper handling will allow the control of software evolution, of systems, and of system instances or views; partial changes or updates to specific parts of a system; and synchronous communications between heterogenous components or even systems. In fact, it is our tool to create synchronous Cartesian systems, essential for context-aware distributed systems. The implementation of a Cartesian distributed system may rely on the behavior of several subsystems, all running on an internal clock necessarily infinitely faster than the external one, since a bunch of tasks in a subsystem, corresponds to one tick of the system. These subsystems all run with respect to a shared context called an æther, which facilitates communication by broadcasting between systems at possibly different levels. The æther in this case is an active context.