Briefly reviews early work in documentation and text processing, and predictions that were made about the creative role of computers in information retrieval. An attempt is made to explain why these predictions were not fulfilled and conclusions are drawn regarding the limits of computer power in text retrieval applications. (Author/CLB)