Article ID Journal Published Year Pages File Type
422658 Electronic Notes in Theoretical Computer Science 2007 18 Pages PDF
Abstract

This paper describes a model-checking based methodology to detect systematic errors commonly made by non-expert users. The human and computer components of the systems are modelled separately. The human component consists of a general model of the user's cognitively plausible behaviour, which can be then refined into specific instances of behaviour that reflect relevant aspects of users' personalities and skills. We consider, as a case study, a formal model of an online interactive tool that enables conference attendees to share thoughts and reactions and select matching attendees to start communication with. Starting from the initial system design, a model-checking technique is used to highlight system vulnerabilities that arise from interactions with non-expert users and may lead to security violations. The results of the analysis are exploited to improve the design by introducing safeguards that reduce or even prevent security violations.

Related Topics
Physical Sciences and Engineering Computer Science Computational Theory and Mathematics