Many claims have been made for the benefits of software reuse, in terms of enhanced quality as well as productivity. Widely reused objects are supposed to possess several desirable characteristics, such as reliability and flexibility as well as efficiency. But reusable objects may be used for unpredicted purposes in unpredicted contexts. Furthermore, in an open distributed world, technical, geographical, organizational and other boundaries, as well as significant time lags, may separate the software developer from the publisher, and the software librarian from the user. From these premises we argue that responsibility for the quality of reusable software artefacts cannot be taken by the developer (or development organization) alone, but must be shared between developer and other agents. Recent work in enterprise modelling for open distributed processing has led to new techniques for modelling responsibilities across organizational boundaries, and these techniques are introduced here as a way of determining and clarifying effective structures for the quality assurance of reusable business objects.