paul2@dcs.exeter.ac.uk (Paul Day) (06/17/91)
been watching the dictonary discusson with interest, and nobody has mentioned (yet) the problems in restricting the language set used inside the dictionary definitions (similar to LDOCE). Restricting the language for definitions to a few 1000 'primitives' would enable the definitions to be used by nlp systems to 'learn' new words. There is also a lot of data (i.e. pronounciation, semantic & syntactic) inside the avergae dictionary - how will that be created ?? Paul Day -- = Paul Day, Centre for Connection Science, Computer Science Dept, = = University of Exeter, Exeter, Devon, England. = = BITNET: paul2@dcs.ex.ac.uk@ukacrl | JANET: paul2@uk.ac.ex.dcs = ========= I didn't do it, nobody saw me do it, you can't prove anything =======