Domain-specific languages and type systems

My primary research interests include functional programming languages, embedded domain-specific languages, type systems, and combinatorics, and ways that these topics can contribute to other areas of computer science. More broadly, running through all of my research are three major themes which shape and direct my efforts: • I am highly motivated by beauty in its many forms. I learn new things because they are beautiful; I strive for beauty in all my creative output; I am motivated to communicate an appreciation for beauty in my teaching. Beauty also correlates with generality: the most beautiful solutions are the ones that have cross-discipline applicability. • I love the process of designing beautiful, coherent visions and then following through on the details to make those visions into reality, which I call architecting. I am not interested in theory just for theory's sake, nor in just hacking out details, but in a harmonious synthesis of the two. • Architecting and appreciating beauty always happen in the context of community. I love communicating ideas, through both teaching and writing, and see communication and collaboration as a central and necessary aspect of my research. Good research is not a solitary pursuit, but a dialogue among researchers—I am always on the lookout for other researchers and students to collaborate with. In what follows, I describe several current areas of research, highlighting the ways they fit into the above themes, and elaborate on planned future directions for my research. Almost any domain can benefit greatly from the principled design of domain-specific languages (DSLs) and type systems. Such languages enable more economical communication of problems and solutions in the domain, and abstracting away domain-irrelevant detail enables higher-level thinking and new insights. Moreover, the process of constructing domain-specific languages itself often sheds new light on the domain under consideration, since it exposes fundamental questions about the meaning of entities and operations in the domain. Type systems play an important role in the design of expressive DSLs. Sme of my earlier work was concerned with extending Haskell's type system to allow " promotion " of values to types [4], a feature particularly useful for DSLs embedded in Haskell. That work has made a big impact in the functional programming world, garnering nearly 50 citations over the past two years. My current work on embedded domain-specific languages focuses on the areas of graphics and animation. For the past six years I have …