Salton award lecture information retrieval as engineering science

Abstract

This paper gives my personal view on the field of information retrieval (IR), by first presenting my own definition of the field, and then pointing out why IR isn't an engineering science yet. For the latter, IR lacks appropriate theoretic foundations, which would not only give us a better understanding of current systems, but also would provide a generally valid basis for building systems, and enable us to make predictons about the performance of these systems. For reaching this goal, we need a well-balanced interplay between theory building, hypothesis generation and testing via experimentation.