Context-aware computing has increasingly gained the attention of the research community because, as it is the case with human interactions, context information provides the background against which it is possible to more accurately interpret communicative acts without the need to explicitly state everything that might be relevant. If, within an agent negotiation for buying some specific service, the service provider says “
the price is 20 Euros
”, the receiver would not be capable of fully interpreting the meaning of the message without using the context created by the whole conversation. Context information provides the basis for more efficient information processing mechanisms due to the possibility of discarding irrelevant information in early stages of information processing. For instance, if some patient’s personal assistance agent is looking for a service that would sell him or her a specific medicine and deliver it in the patient’s home, this would be achieved through the creation of a compound service consisting of an on-line pharmacy and a medicine transportation service. Using context information about the patient’s location, the service composition process may discard service providers located far away from the client and create the compound service considering only a very small number of all existing services of the relevant categories. Context information also enables better adapted behavior since, being context-aware, it may be more directed towards clients requirements in the circumstances of the interaction.