Hier komt nog een Graphic voor
Hier komt nog een Banner Text voor
Translation memories and tools
Give us the tools and we will finish the job.
Winston Churchill, 1874 – 1965.
Nearly every professional translator uses special translation software that is often referred to as a “CAT tool”. “CAT” here stands for “Computer-Assisted Translation” (or “Computer-Aided Translation”). The next few paragraphs give a brief explanation of the way in which this software works.
The first thing that a CAT tool does is analyse the translatable text (either of one file, or of all the files in a translation project) in terms of the total number of words that must be translated. The second thing a CAT tool does is to break up the total text corpus (so, again, one file or all the files in a translation project) into smaller translatable units called segments. This activity is called segmentation. Segmentation is done on the basis of so-called segmentation rules which determine where the one segment ends and the next segment starts.
One this segmentation is complete, the translator can start translating the segments one by one. Every time the translator has completed the translation of a segment, the original sentence and its translation are together stored in what is called a translation memory. Now, these two bits of text (source segment and translated segment) become a translation unit. As time goes by, all these translation units make the translation memory grow bigger and bigger.
And this is the beauty of it. Because, when for example a manual is updated and needs to be translated again, all the translator then has to do is perform an analysis of the new version of the manual and have it segmented. But this time, the translator can do so on the basis of the translation units stored in the translation memory. There will be a lot of segments that have not changed, so all these translation can be reused. The analysis of the segments will now include so-called 100% matches: segments whose translations from the translation memory can be used again and only need to be checked for correctness and completeness. In other cases, for example when only one word in a segment has changed, the analysis will indicate that there is an 84% match in comparison to the original segment plus translation.
A second major component of a CAT tool is a so-called termbase. Like a translation memory, a termbase is also a database component. But in the case of a termbase, the translator does not store translation units but terminology units. The use of terminology in this way enhances consistency and quality, and saves time, as well, as certain CAT tools automatically insert terms from the termbase into the segment that is up for translation. Moreover, after a certain period of time (but depending on the terminology activities of the translator), the translator will have a very valuable asset, namely the bilingual (or sometimes even trilingual) terminology for a certain domain or client.
The above is just a brief and general overview of translation tools and memories. Many tools have additional functions and features that help the translator perform everything that needs to be done. To describe everything in detail would be beyond the scope of this page.
Nico van de Water Linguistic Services currently holds licences for the following CAT tools:
- Déjà Vu X and Déjà Vu X2 (www.atril.com)
- SDL Trados® Suite 2007 Freelance (Translator’s Workbench™ and SDL MultiTerm® 2007 Desktop) and SDL Trados Studio™ 2009 (www.sdl.com/en/products-and-solutions/products/default.asp)
- memoQ (http://kilgray.com/products/memoq)
- Heartsome Studio Enterprise, with update to R6 (www.heartsome.net/EN/home.html)