Positive Data Languages
Abstract
Positive data languages are languages over an infinite alphabet closed under possibly non-injective renamings of data values. Informally, they model properties of data words expressible by assertions about equality, but not inequality, of data values occurring in the word. We investigate the class of positive data languages recognizable by nondeterministic orbit-finite nominal automata, an abstract form of register automata introduced by Boja\'nczyk, Klin, and Lasota. As our main contribution we provide a number of equivalent characterizations of that class in terms of positive register automata, monadic second-order logic with positive equality tests, and finitely presentable nondeterministic automata in the categories of nominal renaming sets and of presheaves over finite sets.
Cite
@article{arxiv.2304.12947,
title = {Positive Data Languages},
author = {Florian Frank and Stefan Milius and Henning Urbat},
journal= {arXiv preprint arXiv:2304.12947},
year = {2023}
}