Unification of Hypergraph $\lambda $-Terms - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2017

Unification of Hypergraph $\lambda $-Terms

Alimujiang Yasen
  • Function : Author
  • PersonId : 1030360
Kazunori Ueda
  • Function : Author
  • PersonId : 1030361

Abstract

We developed a technique for modeling formal systems involving name binding in a modeling language based on hypergraph rewriting. A hypergraph consists of graph nodes, edges with two endpoints and edges with multiple endpoints. The idea is that hypergraphs allow us to represent terms containing bindings and that our notion of a graph type keeps bound variables distinct throughout rewriting steps. We previously encoded the untyped $\lambda $-calculus and the evaluation and type checking of System $F_\texttt {<:}$, but the encoding of System $F_\texttt {<:}$ type inference requires a unification algorithm. We studied and successfully implemented a unification algorithm modulo $\alpha $-equivalence for hypergraphs representing untyped $\lambda $-terms. The unification algorithm turned out to be similar to nominal unification despite the fact that our approach and nominal approach to name binding are very different. However, some basic properties of our framework are easier to establish compared to the ones in nominal unification. We believe this indicates that hypergraphs provide a nice framework for encoding formal systems involving binders and unification modulo $\alpha $-equivalence.
Fichier principal
Vignette du fichier
440117_1_En_9_Chapter.pdf (249.15 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01760636 , version 1 (06-04-2018)

Licence

Attribution

Identifiers

Cite

Alimujiang Yasen, Kazunori Ueda. Unification of Hypergraph $\lambda $-Terms. 2nd International Conference on Topics in Theoretical Computer Science (TTCS), Sep 2017, Tehran, Iran. pp.106-124, ⟨10.1007/978-3-319-68953-1_9⟩. ⟨hal-01760636⟩
73 View
82 Download

Altmetric

Share

Gmail Facebook X LinkedIn More