Regular expression containment as a proof search problem
Abstract
Regular expression containment has recently enjoyed a re- newed interest as a type-theoretic problem. By viewing regular expres- sions as types, it is possible to view containments as functions that coerce one regular expression into another. Containment of a given regular ex- pression in another one is therefore a coercion inference problem. It is not straightforward how to apply here generic proof-theoretic approaches such as those based on the MALL with the least fixpoint. We propose an alternative approach, still a purely syntactic one, to decide contain- ment in reasonable time. We employ the partial derivative construction that, for a regular expression E, yields a syntactic NFA recognising the language of E; and, for a containment E ≤ F , it yields a simulation of E by F , with the empty simulation representing the absence of coercion. We are currently working on a proof of completeness of this optimised coercion search algorithm.
Origin : Files produced by the author(s)
Loading...