A translation memory is a database that stores translated text segments. This means that each segment (in most cases a sentence is a segment, but a segment can also be a phrase) is stored in its original version and translated.
What’s the purpose of a translation memory?
In the first place, the purpose of working with a translation memory is to speed up the translation process. It’s especially useful for long, technical or scientific translations in any field and in almost any language pair.
When not to use a Translation Memory
It’s not recommended to use a translation memory when you’re translating literary text. In this case, segmentation plays against the tools that a translator with this kind of expertise has. Literary translation is characterized by great flexibility, especially at the structural level, and the ability to “play” with the text. It is common to join sentences or, on the contrary, to need to cut out a paragraph: what matters is to be able to make the text as flexible as possible and for this purpose, working with a computer-assisted translation tool is not ideal.
A translation memory is not a software per se, but you could say it’s a “bilingual file” to keep within a software (that is, a tool for computer-assisted translation) in order for it to remember how sentences (segments) and words are translated.
Both the translator and the client can benefit from using software to manage translation memories. Some of these benefits include:
- Ensuring that the document is fully translated
- Being consistent throughout the translation
- No need to worry as much about the format
- Significant savings in cost and time, especially in large projects