Perhaps, the most important aspect of modern translation industry is using of so-called CAT (computer assisted translation) tools that integrate translation, quality control and terminology check functions. These programs have nothing to do with automatic translation tools; actually they DO NOT TRANSLATE but rather help a translator to make the target text as consistent and thus as good as possible. It is always a human who makes the translation. CAT tools just give a translator some useful “tips” that help avoid mistakes and significantly increase translation quality.

How does it work?

The heart of most translation tools is so-called translation memory, a database that stores original (source) text segments and their respective translations (target). When a translator comes to a new segment the program automatically searches its memory to find whether any identical or similar segments have been translated earlier or not. Then the program displays search results (matching segments) and their translations in a separate window. It also shows percentage of similarity between the source text to be translated and the one from the translation memory.

Let us consider the following situation. A customer wants an agency to translate 10 similar product specifications. When translation of the first one is complete all translated segments are stored in the memory. When a translator starts working on the rest the program shows him translation of the matching segments. Particularly in this case 9 specifications can be processed within the same time frame as the first one.