ABSTRACT
This paper is intended to provide an analytical background for those seeking effective text retrieval systems, and more specifically for those advocating the application of techniques drawn from artificial intelligence (AI) and natural language processing (NLP) for this purpose. A new wind is blowing through the world of information retrieval, and it seems that some of the apparent limitations of existing methods for characterizing and retrieving text-based information can be overcome. These existing methods refer primarily to information retrieval in the sense of document retrieval. Much of what is proposed falls, explicitly or implicitly, under this broad heading; and it is also useful to approach other forms of information retrieval from document retrieval, to make their distinctive properties and implications clear.
