Machine translation refers to a piece of text which have been written again from one language into another language by a machine. Nowadays the term ‘machine’ is usually the same as computer in this context. Because, unlike Thomas Watson, the chairman of IBM at that time, famously stated in 1943, computers have become the most important tools in many aspects of human life and because machine translation is relatively new term, it is useful to explain the concept a little more.
Practically speaking, machine translations are made with the help of an existing translation memory. This kind of memory can be a lot of things, but basically it is a file containing the same text written in more than one languages. For example any electronic dictionary can be used as a translation memory. Machine translator then uses the memory as a source for translations. This evidently leads to the situation where the more comprehensive the translation memory the better the translation quality.
Computers can perform machine translation with or without an Internet connection. Perhaps the most known online machine translation provider is Google with its free Google Translate -service. Online machine translation can utilize online translation memory while offline machine translators must have their own memory. Both of these methods have their pros and counts: online memory can be very huge and automatically updated, but at the same time they are public and no one can really control them. Offline memories on the other hand can be strictly controlled by the user but they have to be manually collected and updated.
This was a very short overview of machine translation. If you want to take a deeper look at the world of machine translation, you can read the article in Wikipedia for example.
How would you define the term ‘machine translation’?