Papers
Topics
Authors
Recent
2000 character limit reached

Game Semantics: Easy as Pi

Published 10 Nov 2020 in cs.PL and cs.LO | (2011.05248v1)

Abstract: Game semantics has proven to be a robust method to give compositional semantics for a variety of higher-order programming languages. However, due to the complexity of most game models, game semantics has remained unapproachable for non-experts. In this paper, we aim at making game semantics more accessible by viewing it as a syntactic translation into a session typed pi-calculus, referred to as metalanguage, followed by a semantics interpretation of the metalanguage into a particular game model. The syntactic translation can be defined for a wide range of programming languages without knowledge of the particular game model used. Simple reasoning on the model (soundness, and adequacy) can be done at the level of the metalanguage, escaping tedious technical proofs usually found in game semantics. We call this methodology programming game semantics. We design a metalanguage (PiDiLL) inspired from Differential Linear Logic (DiLL), which is concise but expressive enough to support features required by concurrent game semantics. We then demonstrate our methodology by yielding the first causal, non-angelic and interactive game model of CML, a higher-order call-by-value language with shared memory concurrency. We translate CML into PiDiLL and show that the translation is adequate. We give a causal and non-angelic game semantics model using event structures, which supports a simple semantics interpretation of PiDiLL. Combining both of these results, we obtain the first interactive model of a concurrent language of this expressivity which is adequate with respect to the standard weak bisimulation, and fully abstract for the contextual equivalence on second-order terms. We have implemented a prototype which can explore the generated causal object from a subset of OCaml.

Citations (6)

Summary

Paper to Video (Beta)

Whiteboard

No one has generated a whiteboard explanation for this paper yet.

Open Problems

We haven't generated a list of open problems mentioned in this paper yet.

Continue Learning

We haven't generated follow-up questions for this paper yet.

Collections

Sign up for free to add this paper to one or more collections.