For a long time, computer specialists and translators saw each other as competitors or even enemies. It was widely believed that computer programs would eventually be cheaper, faster, and better than human translators so that training human translators would be superfluous. This belief was widespread among computer scientists as well as among translators. After more than half a century of research in machine translation we can confidently say that this belief was wrong. It was based on mutual misunderstanding. Computer scientists underestimated the complexity of the translation task and translators overestimated the capacities of ‘intelligent’ computer programs.

It is only since the early 1990s that a reorientation of the efforts at the interface of computing and translation can be observed. Computational linguists started asking what translators would like that computer programs would do for them. This was the starting point for a new generation of computer tools for translators.

Computer Environments for Translators

Termbases and translation memories are the core of computational support tools for translators. Around them, different software packages organize a variety of further tools and options that facilitate the work of a translator. They require some special training to get used to, but they are designed for use by translators, not by computer specialists. They increase the productivity of the translator as well as the quality of the translation by supporting the components of the job that translators like the least: repetitive formulations and technical terms. Fortunately, these are the components computers are best able to handle.