This volume brings together in revised form 17 of the 45 papers presented at the International Workshop on Parsing Technologies hosted by Carnegie Mellon University in August 1989. A companion volume, Generalized LR Parsing, (reviewed by Lopez, this issue) contains a further topic-based selection of papers from the same workshop.
In a short introductory chapter to the volume, the editor defines parsing technologies as a research area concentrating on the non-linguistic or linguistically uninteresting problems which must be solved in practical NLP applications. He takes both the restrictions on the format of date expressions in English (*on 15th July) and also, surprisingly, idioms, to be non-linguistic problems. The area of research also includes the implementation of linguistic theories and formalisms, research on their formal properties, on parsing algorithms and on probabilistic methods. The introduction argues that this sort of research is essential both in view of the practical goals of NLP (real applications) but also in order to bring down novel linguistic theories and insights into those computational applications.
The papers in this volume address a very wide range of topics. A number of papers are "linguistically committed" in the sense that they address issues in parsing within a particular theoretical linguistic framework (e.g. Fong and Berwick) or indeed describe particular parsers (Jensen). At the other end of the spectrum is the contribution by Tomita, which discusses the 2-dimensional parsing of input text using a 2-dimensional context-free grammar (CFG). The interpretation of layout information is crucial to the understanding of a range of texts including title pages, announcements, diagrams and figures (e.g. the failure to understand this sort of information is a frequent cause of poor results from the use of optical character readers). In a 2d-CFG the symbols on the right-hand sides of rules can