Section: New Results

Language interoperability: ML and a Linear language

Participant : Gabriel Scherer.

In a programming system where programs are created in one programming language, we consider the addition of another programming language that interoperates with the first – and the reimplementation of some library/system functions in this new language. This can increase expressivity, but it could also break some assumptions made by programmers. Typically, adding a bridge to C or assembly code can introduce memory-unsafe code in a previously-safe system. In [18], we formalize a notion of “graceful” interoperability between two languages in this setting, determined by full abstraction, that is, preservation of equational reasoning. We instantiate this general idea by extending ML with an advanced expert language with linear types and linear mutable cells.