WG211/M19Walkingshaw

Eric Walkingshaw: Toward a Variational Programming Language

Variational programming is an emerging paradigm for representing, analyzing, and computing with variation in code and data. It is useful whenever a program must be run on many related inputs, such as during speculative analysis, or when many similar programs must be executed, such as when testing software product lines or computing with multiple privacy policies. Variational programming enables scaling computation to run over large numbers of similar variants by making variation explicit in data and code, preserving variation during execution, and exploiting similarities across variants through sharing.

In this talk, I will present current and proposed work toward the design and implementation of a general-purpose language for variational programming. The goal of this language is to: (1) advance research on theoretical foundations of variational programming, (2) support research on variational data structures and algorithms, which are needed as basic building blocks for variational programs, (3) promote variational programming as a general and widely-applicable problem solving strategy, and (4) make variational programming practically accessible to programmers and researchers in the many application domains where it can be put to use.