1998•Formal Aspects of ComputingOpen access

Unification of Four Versions of Program Semantics

Ingrid Rewitzky, Chris Brink

Open full text 2 citations

Abstract

Abstract. Computing Science is a new subject, and we have not yet achieved the unification of theories that should support a proper understanding of its structure . CAR Hoare and He Jifeng, 1998. In this paper we use Priestley duality and Jónsson/Tarski duality to translate between four versions of program semantics: the relational model, predicate transformer semantics, information systems, and powerdomains. Our point of entry is the relational model, a kind of Kripke semantics, in which programs are thought of as input-output relations over a structured state space. Specifically, we present the state space as a certain kind of Priestley space, and programs as certain structure-preserving relations over such a space. We then derive, in circular fashion, a predicate transformer semantics from the relational model, an information system from the predicate transformer semantics, a powerdomain from the information system, and the original relational model back again from the powerdomain. The information system is also shown to be related to Hoare logic. To clarify the intuition behind this approach we present a case study, which is a ‘Priestley version’ of the so-called universal domain due to Plotkin, and we explicate various ideas about properties of programs and predicates in terms of this case study.

Open-access reader

About this research paper

What this paper is about

Abstract. Computing Science is a new subject, and we have not yet achieved the unification of theories that should support a proper understanding of its structure . CAR Hoare and He Jifeng, 1998. In this paper we use Priestley duality and Jónsson/Tarski duality to translate between four versions of program semantics: the relational model, predicate transformer semantics, information systems, and powerdomains. Our point of entry is the relational model, a kind of Kripke semantics, in which programs are thought of as input-output relations over a structured state space. Specifically, we present the state space as a certain kind of Priestley space, and programs as certain structure-preserving relations over such a space. We then derive, in circular fashion, a predicate transformer semantics from the relational model, an information system from the predicate transformer semantics, a powerdomain from the information system, and the original relational model back again from the powerdomain. The information system is also shown to be related to Hoare logic. To clarify the intuition behind this approach we present a case study, which is a ‘Priestley version’ of the so-called universal domain due to Plotkin, and we explicate various ideas about properties of programs and predicates in terms of this case study.

Why it matters

OpenAlex reports 2 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Abstract. Computing Science is a new subject, and we have not yet achieved the unification of theories that should support a proper understanding of its structure . CAR Hoare and He Jifeng, 1998. In this paper we use Priestley duality and Jónsson/Tarski duality to translate between four versions of program semantics: the relational model, predicate transformer semantics, information systems, and powerdomains. Our point of entry is the relational model, a kind of Kripke semantics, in which programs are thought of as input-output relations over a structured state space. Specifically, we present the state space as a certain kind of Priestley space, and programs as certain structure-preserving relations over such a space. We then derive, in circular fashion, a predicate transformer semantics from the relational model, an information system from the predicate transformer semantics, a powerdomain from the information system, and the original relational model back again from the powerdomain. The information system is also shown to be related to Hoare logic. To clarify the intuition behind this approach we present a case study, which is a ‘Priestley version’ of the so-called universal domain due to Plotkin, and we explicate various ideas about properties of programs and predicates in terms of this case study.

Key concepts: Unification, Computer science, Predicate (mathematical logic), Principle of compositionality, Predicate transformer semantics, Programming language, Denotational semantics, Operational semantics

Related papers

Back to paper searchBrowse research topicsOriginal source
Unification of Four Versions of Program Semantics — Research Paper | ScholarLens