© 1989 by Association for Literary & Linguistic Computing
Articles |
Program-Generator for OCP in a Language with Heavy Morphology
Technion-lsrael Institute of Technology Haifa, Israel
Uzzi Oman, Faculty of Computer Science Tech-nion-lsrael Institute of Technology Haifa, Israel Electronic mail ornan@techsel bitnet, ornan@sel technion ac il csnet
The Oxford Concordance Program (OCP) may collect various forms under a preferred entry. In order to activate this option, the user should write down all the related forms in a special line. This task is rather slow and tiresome, especially when the text is in a language with heavy morphology We replaced the boring task of writing forms with a program generator, which creates OCP programs. Once such a generator has been prepared, it may be used in any future program. Also, the preparation of it is rather creative, thus not annoying. The grammatical data consists of a lexicon, with a number annexed to each entry. This number is a pointer directing to the other part of the grammatical data, ordered instructions for making changes in the entry, then generating all its related forms.
Our program (written in PASCAL) first writes down the fixed parts of an OCP program, then it turns to the lexicon, writes down its first entry, moves to the appropriate instruction (according to the pointer), and generates all forms which are related to the entry, then it turns to the next entry of the lexicon and so on until the last desired one. The program now turns to the concluding lines of OCP and copies them. Preparing the lexicon and writing down morphological instructions for generating related forms is an investment which might have been done for the sake of the Hebrew grammar, not necessarily for OCP programs. They may be used for all kinds of language processing as well as for teaching purposes. In any case, once you have prepared them, you may use them as they are in any further use of OCP. Another conclusion is that such a device may be used also for other languages