Around 1970 a system called Systran was developed at Dayton, Ohio. Its purpose was to translate Russian into English by computer. Since then other systems have been developed which use computers to translate from one language into another. This is called machine translation, or MT. so far, only very large organizations have had enough money to spend on this kind of development. Such organizations are also aware of an urgent need for MT. in 1985 a company was spending around $ 300 million a year on translation. Translation by a good human translator would at that time cost about $10 per page. Each page would take the same translator about an hour to translate. Translation is a costly and time-consuming business. Obviously, much time and money could be saved by MT in the fields of diplomacy, science and technology and international business.
At present many European universities are working on a means of translation seven major European languages by computer. Such projects are not available for ordinary commercial users. It is, however, possible to hire MT services for specific jobs.
After these projects the next stage, which is now under way, is to develop a system which can be used on microcomputers in an ordinary office. Such systems would work in a similar way to the spelling checker, available in word processor programs, which has a dictionary of 40,000 -90,000 word. A foreign language dictionary which would work in the same way is a possible way of producing word for word translation.
However, there are limitations to what MT can do. Systran translates correctly only eight words out of ten. Simpler systems have an even worse record. Therefore, MT texts still require editing by a human translator.
However, with the large amount of the translation work being done by the computer, time is saved by this method.