Languages for Secure Multiparty Computation and Towards Strongly Typed Macros

We show that it is feasible and useful to create programming languages with strong security guarantees for secure multiparty computation. We have designed and implemented the Secure Multiparty Computation Language (SMCL), which is a domain-specific programming language for secure multiparty computation. SMCL allows programmers to write programs using secure multiparty computation without expert knowledge on how to design and implement cryptographic protocols. We have proven that programs written in SMCL are immune to a broad range of security threads and confidential information may only be revealed in specific parts of a program, designated by the programmer. We demonstrate the usefulness of SMCL by reporting on how an SMCL program contributed to the first large-scale practical application of secure multiparty computation. Based on our experiences with SMCL we have designed a successor called PySMCL which is a domain-specific language embedded in Python, and will provide even better security guarantees than SMCL. We also show that syntactic abstraction can enable programming languages to evolve over time by adding new libraries with syntax and semantics as they become needed. This is an important feature because it is difficult to anticipate which language concepts becomes needed over the lifetime of a programming language. We have designed and implemented a hygienic syntactic abstraction system for the Fortress programming language based on parsing expression grammars. The system allows syntactic extensions of Fortress to be specified in a modular fashion and added in libraries as needed. The system also allows us to not only support new language extensions, but also to move many constructs of the core language into libraries. We have designed a type system which extend the guarantees provided by the Fortress type system to the combined language of Fortress and our system for syntactic abstraction.