This thesis describes the design and implementation of a software agent, Arbitrator Agent, which supports online negotiation of privacy contracts for e-commerce. It introduces a usage concept consisting of a triple of P3P-defined elements (purpose, recipient, retention) upon which both users and businesses can usefully negotiate. Furthermore, a negotiation terminology model has been proposed, which supports subdivision of complex user rules into sub-preferences and business data requests into smaller sub-requests. The disjoint data set types in the P3P schema is exploited for fast determination of whether negotiation may be useful.; An ontological representation of the P3P data schema, and distance functions applied to sub-requests and sub-rules are utilized in a novel approach to finding relevant substitute data for counteroffers within a negotiation session. The implementation approach uses a hybrid of semantic web ontology language and conventional data structures to support our contextual negotiation application. (Abstract shortened by UMI.)... |