Automated abstracting
Christine A. Montgomery · 1981
The automatic production of extract-type abstracts was one of the very early ideas for the application of computers to the processing of natural language text. In an early article, Hans Peter Luhn of IBM discussed the possibility of producing extracts of scientific and technical materials by frequency counts of occurrences of content words [Luhn 1968]. Unfortunately, as in the case of machine translation, the linguistic, statistical, Software and hardware technology were not mature enough to validate—or invalidate—the claims that were made for various methodologies by the pioneers in the field. Thus activities in these two areas were characterized by a great deal of arm-waving, claims and counter-claims, and generally trivial experiments which validated or invalidated nothing in particular.