Before we start discussing this topic, we would like to clarify first what machine translation (computer assisted translation) actually is. Machine translation is a product of computational linguistics, a field which explores ways of using computer software to translate texts or speech from one language into another.
The first such software was created back in 1954. As part of the so-called “Georgetown – IBM“ experiment, over sixty sentences were translated from Russian into English. However, even though the experiment was a success at the time, problems appeared when certain terms were translated.
In 1959, IBM developed translation software Mark I, and by 1963, Mark II had already been translating up to 5000 words per hour. Technological advancement and the war in Vietnam lead to the development of IBM 360/67, software which could translate between 80,000 and 100,000 words per hour, from English into Vietnamese.
Today, there are many types of software which are used for translation purposes. One person can translate a maximum of 500 words in one hour, while one computer program is capable of translating up to 4,000 words per minute. Software used for translating can also store translated documents in their memories and reuse them if necessary.
However, not a single translation, created by a machine, can be characterized as perfect. In this respect, the human factor is still an irreplaceable ingredient of any quality translation.
* Computer Assisted Translation