A theory of contexts in information bases

Abstract

Although semantic data models provide expressive conceptual modeling mechanisms, they do not support context, i.e. providing controlled partial information on conceptual entities by viewing them from different viewpoints or in different situations. In this paper, we present a model for representing contexts in information bases along with a set of operations for manipulating contexts. These operations support context creation, update, copy, union, intersection, and difference. In particular, our operations of context union, intersection, and difference are different from those of set theory as they take into account the notion of context. However, they also satisfy the important properties of commutativity, associativity, and distributivity. Our model contributes to the efficient handling of information, especially in distributed, cooperative environments, as it enables (i) representing (possibly overlapping) partitions of an information base; (ii) partial representations of objects, (iii) flexible naming (e.g. relative names, synonyms and homonyms), (iv) focusing attention, and (v) combining and comparing different partial representations. This work advances towards the development of a formal framework intended to clarify several theoretical and practical issues related to the notion of context. The use of context in a cooperative environment is illustrated through a detailed example.