Conal Elliott

Distinguished Scientist at Target
 conal.net
 @conal

Conal Elliott has been working (and playing) in functional programming for more than 35 years. He especially enjoys applying semantic elegance and rigor to library design and optimized implementation. He invented the paradigm now known as “functional reactive programming” in the early 1990s, and then pioneered compilation techniques for high-performance, high-level embedded domain-specific languages, with applications including 2D and 3D computer graphics. The latter work included the first compilation of Haskell programs to GPU code, while maintaining precise and simple semantics and powerful composability, as well a high degree of optimization.

Conal earned a BA in math with honors from the College of Creative Studies at UC Santa Barbara in 1982 and a PhD in Computer Science from Carnegie Mellon University in 1990. He is currently working as distinguished scientist at Target. Previously, we worked at Tabula Inc on chip specification and compiling Haskell to hardware for massively parallel execution until their closure in early 2015. Before Tabula, his positions included Architect at Sun Microsystems and Researcher in the Microsoft Research graphics group.

 

YOW! Lambda Jam 2017 Sydney

Teaching New Tricks to Old Programs

KEYNOTE

Many useful operations are well-defined on functions but are not computable, e.g., root-finding, optimization, exact differentiation and integration, and efficient, incremental evaluation. Sometimes these problems can be solved by means of a domain-specific embedded language (DSEL) with an implementation that maintains extra information. With extra effort, these implementations can be quite efficient, but at the cost of duplicating work of the host language compiler. Although overloading can hide some of the required change of vocabulary, the illusion is imperfect, and so code must be rewritten for the new DSLs with sometimes awkward or surprising results.

This talk presents an alternative to DSELs, giving new interpretations to existing functional programs. The implementation is a plugin for GHC—a popular, high-quality Haskell compiler—and works by translating to a well-known, more easily generalizable form. Each new interpretation is simply a new type and a collection of class instances for it, written in standard Haskell, with no exposure to compiler internals. To get a feel for the breadth of this technique, we’ll look at interpretations including hardware circuits, automatic differentiation, incremental evaluation, and interval analysis.