At the beginning of this paper, I said that the information world in 1945 was ready and waiting for the computer. In the ensuing fifty years, the computer has of course become the major tool for the practice of IR. But more importantly, it has been instrumental in changing fundamentally our perception of the nature of information retrieval. Together with the development of theories and models, and of the perception of IR as a field for empirical investigation and experiment, the mere existence of the tool has served to alter our ways of thinking about the problem, almost out of all recognition.

This process is by no means complete. All three reasons remain powerful agents for change, individually and in combination. For example, we have scarcely begun to get to grips with the effective empirical study of highly interactive systems (such as graphic user interfaces) and their use in retrieval. And while there is not now (and may never be) any overall theory of information retrieval, new models and theories continue to contribute to our understanding.

Information retrieval is a real-life problem, and as such has generated many practical papers discussing operational systems and services. It is also an area of exciting intellectual endeavour which is worthy of, and has sometimes attracted, real intellectual power. Reading back over old issues of the Journal, as an active participant myself, I was sometimes inevitably annoyed at what, with hindsight, I perceive as naïvety or worse. I was also, however, often stimulated by the ideas, arguments, ingenuity, and foresight that I found. We have come a long way since 1945, but the journey has been as interesting in retrospect as it was at the time, and as the field is now. I firmly believe that the next fifty years will be just as exciting.