Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper addresses the complexities of proving the Endpoint Projection (EPP) theorem in choreographic programming by introducing a novel semantics that aligns the global view of choreographies with the local view of processes. The authors develop a new preorder relation that extends bisimulation to account for knowledge propagation among distributed processes, enabling a modular proof of EPP. This approach not only simplifies the proof but also enhances understanding of choreographic programming theory, making it more accessible for researchers and practitioners alike.
A modular proof of the Endpoint Projection theorem reveals deeper insights into choreographic programming by bridging the gap between global and local perspectives.
The Endpoint Projection (EPP) theorem is a cornerstone of choreographic programming. It states that every choreography can be projected to a network of processes that correctly implements it. Proving EPP is notoriously difficult, and existing proofs are complex and non-modular because of the mismatch between the global view of choreographies and the local view of processes. In this article, we show how to reconcile this mismatch by designing a new semantics for choreographies that is built on the local view of processes, as well as a new preorder relation between choreographies and networks that extends bisimulation to deal with the propagation of knowledge of choice among distributed processes. As a result, we can give a modular proof of EPP, which is conceptually simpler than existing ones and also provides better insights on the theory of choreographic programming.