Self-adjusting computation refers to a model of computation, where every program can responds to changes to its data by updating its output automatically. The idea is to represent a computation in such a way that when its data changes, all parts of the computation that depend on that change can be updated efficiently. Realizing this idea requires combining language design (type theory and semantics) and algorithm design and implementation. Applications abound in various areas such as algorithms, compilers, computational biology, computational geometry, machine learning, and scientific computing. Please see our papers for more information.
We have several implementations of the model written in the Standard ML and the C languages.