ABSTRACT
Several decades ago, we opened a book to acquire knowledge, but today we turn on the computer. The Internet, like an encyclopaedia, provides all kinds of information presented in various languages, but we cannot understand it without breaking down the language barrier. Although machine translation (MT) enhances web-based communication by providing automated translation, the quality of MT output is far from satisfactory to users because human languages show high semantic and syntactic complexity. Languages are so diverse and changeable that neither fixed grammatical rules nor sophisticated computing processing mechanisms can produce accurate translations. Despite the failure of fully automatic high-quality machine translation (FAHQMT) that was initially pursued, there have been some successes, such as the Systran system (1976) used by the Commission of the European Communities in Luxemburg and the Taum Météo system (1978) for French-to/from-English translation of Canada’s weather forecast (Somers 2003). As of this writing, these MT tools have been in use for more than 38 years. LOGOS, ALPS, ENGSPAN, METAL, GLOBALINK are also MT success stories (Arnold et al. 1994).
