Multi-language programming systems: a linear experiment

Instead of a monolithic programming language trying to cover all features of interest, some programming systems are designed by combining together simpler languages that cooperate to cover the same feature space. This can improve usability by making each part simpler than the whole, but there is a risk of abstraction leaks from one language to another that would break expectations of the users familiar with only one or some of the involved languages. We formally study this problem by reusing ideas from previous work on multi-language semantics, used to study modular compilation, that suggest a formal definition of what it means for a given language to be usable without leaks: it should embed into the multi-language system in a fully abstract way, that is, its contextual equivalence should be unchanged in the larger system. This strong formal requirement does not hold in most existing systems; is it a attainable goal for language design? As a first experiment, we design a multi-language programming system that combines a simple ML-like language and a very simple linear language with linear state. The goal is to cover a good part of the expressiveness of languages that mix functional programming and linear state (ownership), at only a fraction of the complexity. We prove that the embedding of ML into the multi-language system is fully abstract: functional programmers should not fear abstraction leaks. We show examples of combined programs demonstrating in-place memory updates and typestate-like usage protocols.