Parallel access to an English dictionary

This paper describes how a computer-based dictionary can be accessed in parallel. The application, which is written in Occam-2, uses up to five transputers and the dictionary is split over these. Words input to the system are preprocessed in order to find the relevant section of the dictionary, except in the case of high-frequency words, which are dealt with separately by means of a trie search. Morphological analysis is employed to find the roots of words that appear in different forms from those listed in the dictionary. Results of implementations with and without an index for the dictionary are presented in respect of various inputs consisting of individual words and sentences requiring use of the trie search, the different sections of the dictionary and the morphological analysis.