Also published as: Rene Leermakers
Much of mathematics, and therefore much of computer science, is built on the notion of sets. In this paper it is argued that in parsing theory it is sometimes convenient to replace sets by a related notion, bunches. The replacement is not so much a matter of principle, but helps to create a more concise theory. Advantages of the bunch concept are illustrated by using it in descriptions of a formal semantics for context-free grammars and of functional parsing algorithms.