Reasoning about the safety of information: from logical formalization to operational definition

Abstract

We assume that safety of information stored in a database depends on the reliability of the agents who have performed the insertions in the database. We present a logic S to represent information safety, and to derive answers to standard queries and to safety queries. The design of this logic is based on signaling act theory. Two strong simplifications lead to a logic S" with two modalities to represent explicit beliefs and implicit beliefs. Then, we present an operational view of S" in terms of First Order Logic, with meta predicates, which is implemented by a Prolog meta program. lt is proved that answers derived in S" and computed by the meta program are identical. This property gives a clea.r meaning to computed answers. Content areas: Epistemological foundations, Theorem proving, Logic programming, Multi-agent systems.