Collection classes are ideal candidates for software reuse and therefore have an important role in the development and use of object-oriented class libraries. In this paper we analyze represent-ative collection libraries, pointing out the main problems from the viewpoint of their practical usability in programs. We then identify a small number of independent concepts relevant to the design of libraries of collections and show how an orthogonal treatment of these concepts can solve the problems identified. Because existing object-oriented languages, even newer ones, do not completely support all the relevant concepts (although most of the critical points are well-known), we show to what extent it would be possible to realize our proposals, and what remains as a challenge in the design of future languages.