ISR11 Summary summary Harvard University Gerard Salton Use, reproduction, or publication, in whole or in part, is permitted for any purpose of the United States Government. with the SM[OCRerr]T system which make it possible to obtain improvements in subsequent searches, using feedback information supplied by the users as a result of earlier searches. Evaluation results comparing the effectiveness of some of the automatic analysis and search procedures incorporated into the SM[OCRerr]T system were first published in report ISR-8 in this series, dated December l961[OCRerr]. More extensive evaluation output is included in the present report, summarizing the work performed during the fall of l96[OCRerr] and the first half of 1966. The present report contains work in three main subject areas automatic and semi-automatic dictionary construction, evaluation output based on results obtained by processing four document collections in three subject areas, and iterative search experiments based on user feedback. Section I by G. Salton contains a short report on the present state of the SM[OCRerr]T project, including also a summary of the research proposed for the immediate future. A complete set of operating instructions for the present version of the S[OCRerr]T system is presented in section II by M. Lesk. A study of this section should make it possible to other interested parties to run portions of the SM[OCRerr]T system on different 7O9[OCRerr] installations. Various aspects of the automatic dictionary construction problem described in sections III, IV and VIII of the present report. Section by G. Hochgesang contains a description of a very fast concordance generating program which produces keyword-in-context (}c[OCRerr]c) type output from ordinary text input. This program is used to generate the concor- dances which are later incorporated in the dictionary construction system. are III xiv