Historical Background

In the early 1990s, a new way of machine translation emerged. IBM pioneered statistical translation by trying to develop the first system using bilingual text corpora, i.e. sample translations, to produce new translations for texts that are related to the example translations the source text is compared with. The special feature of Statistical Translation Systems is that syntactic and semantic rules have no longer been necessary.
After pure research had been finished, software manufacturers started to develop practical applications. Since 1994 machine translation software for personal workstations has been affordable for most people. Nowadays, even free translation systems are available and popular. The Google translation tool, for example, is Google's most used service next to its internet search.