Fundamenta Informaticae, 33(4):369-396, April 1998.
This article is the complete version of a paper presented at TLCA '97. A preliminary version appeared as BRICS Report RS-97-39 in December 1997.
Fetch article (PDF, 222k). Go to publications, talks or home page.
The nu-calculus of Pitts and Stark is a typed lambda-calculus, extended with state in the form of dynamically-generated names. These names can be created locally, passed around, and compared with one another. Through the interaction between names and functions, the language can capture notions of scope, visibility and sharing. Originally motivated by the study of references in Standard ML, the nu-calculus has connections to local declarations in general; to the mobile processes of the pi-calculus; and to security protocols in the spi-calculus.
This paper introduces a logic of equations and relations which allows one to reason about expressions of the nu-calculus: this uses a simple representation of the private and public scope of names, and allows straightforward proofs of contextual equivalence (also known as observational, or observable, equivalence). The logic is based on earlier operational techniques, providing the same power but in a much more accessible form. In particular it allows intuitive and direct proofs of all contextual equivalences between first-order functions with local names.
@Article{stark:namerp-fi,
author = {Ian Stark},
title = {Names, Equations, Relations:
Practical Ways to Reason about {\em new}},
journal = {Fundamenta Informaticae},
year = 1998,
month = {April},
volume = 33,
number = 4,
pages = {369--396},
note = {Complete version of \cite{stark:namerp-tlca}},
url = {http://www.inf.ed.ac.uk/~stark/namerp-fi.html}
}