Part of my thesis developed a type theory ("Type-indexed rows") which nicely generalised extensible records and variants. One use of the calculus is to encode XML DTD-like declarations. This idea was explored in the "gedankenexperiment" language XMLambda, and could be scaled-up to Haskell. In this talk I'll take a critical look at this approach, contrast it with the regular-expressions-as-types and modal-logic approaches, and outline some thoughts on how to incorporate these ideas into Haskell.

I completed my PhD at the Oregon Graduate Institute in early 2001, and since then have been at Microsoft Research, Cambridge.