Translation Corpora and Machine Aided Translation
Translation Today · 2009
The modification of the goal from holistic automatic machine translation to domain-specific machine-aided translation has been the outcome of our rude realization of the fact that the development of a fully automatic machine translation system, which will work equally effectively for all kinds of text and for all kinds of natural language, is actually a myth.The long history of failures recorded in last six decades in the act of developing an automatic machine translation system for any two natural languages has eventually forced us to think of developing domain-based machine translation system, which gives us a better assurance of success among related languages.Against this back-drop I have made an attempt here to show how linguistic data and information intelligently extracted from systematically designed translation corpora can lead us a few steps forward towards our goal.The method proposed here will be highly useful in the Indian context if we complete the ground work necessary for achieving success before we venture on the voyage.