It is now a truism that libraries will need to change radically if they are to survive the networked era, and not simply be closed down because they are no longer perceived to be needed, In a world where any information in any form will in principle be able to be sent from anywhere to anywhere 'at the touch of a button' everyone realises that the anticipatory 'just-in-case' role that libraries and their librarians have traditionally performed comes under threat from 'just-in-time' solutions, Yet the global information system within which each individual library is just one small player is extraordinarily complex, This means that it is very difficult for those in charge of the destiny of libraries effectively to plan for the future. It is not that we lack information: every month yet more and more 'relevant' papers, reports, books and so on appear, yet more conferences, seminars, workshops and so on are held: all grappling with one or another aspect of library futures. The problem is summarizing and organizing all of the resulting facts and opinions into a form which 'top management' can and will respond to with confidence, if not even enthusiasm! A classic method which policymakers use in such situations is the creation of a model. This paper develops a model which it is hoped is rich enough to be faithful to the underlying complexity it represents; yet not so rich that it overwhelms the (nonspecialist) decision maker. The paper is in the form of an Experiment - whereby the Director of a Research Library explains to the Library's Board of Trustees, why the Library should not be closed down as had been suggested by the parent institution's Director of Computing. The reader is invited to judge whether the Experiment is a success. And, if not, how he or she would respond to such a challenge!.
