Accountable Trust Decisions: A Semantic Approach

Abstract

This paper is concerned with the question of how to obtain the highest possible assurance on trust policy decisions: when accepting an electronic transaction of substantial value or significant implications, we want to be sure that this did not happen because of a bug in a policy checker. Potential bugs include bugs in parsing documents, in signature checking, in checking trust lists, and in the logical evaluation of the policy. This paper focuses on the latter kind of problems and our idea is to validate the logical steps of the trust decision by another, complementary method. We have implemented this for the Trust Policy Language of the LIGHTest project and we use the completely independently developed FOL theorem prover RP_X as a complementary method.