Measuring Privacy

Tracy Ann Kosa¹, Khalil EI-Khatib², and Stephen Marsh³

¹Faculty of Science,

University of Ontario Institute of Technology,

2000 Simcoe Street, Oshawa,

Ontario, L1H7K4, Canada

TracyAnn.Kosa@uoit.ca

²University of Ontario Institute of Technology,

Oshawa, Canada

Khalil.El-Khatib@uoit.ca

³Communications Research Centre Canada

Ottawa, Canada

steve.marsh@crc.gc.ca

Abstract

There is no unified theory of privacy. Law, political science, economics, sociology and psychology

have thoroughly explored the concepts of privacy, while computer science has attempted to apply

these concepts with varying degrees of success. The study of privacy is often lost in a debate over

values, whether privacy itself is a good thing or a bad thing, and how / when it may be reasonably

invaded. This paper ignores that debate, reasoning that privacy is legislated so the values issue is no

longer relevant, and proposes a theoretical mechanism for measuring privacy using trust as a model

based on the need (briefly examined in Section 3) that knowledge about an individuals state
of privacy is necessary. Presenting 3 different sets of factors (human, computer and data) derived from

multiple disciplines, this work identifies the list of considerations from which a state of privacy may

be derived in any given situation; physical or virtual world. This work proposes an original model

of the states of privacy based on the identifiability of an individual. Representation is a finite state

machine, while the same list of factors can be used to calculate transitions in the machine.

Keywords: Privacy, Trust, Finite State Machine

Journal of Internet Services and Information Security (JISIS), 1(4): 60-73, November 2011 [pdf]