Wouldn’t it be great if you didn’t have to continually translate the same phrases over and over again?
Imagine how much time and money it would save as you seek to get your product or service to market in as quick a time as possible as cheaply as possible. This is where Translation Memory (TM) comes in, as the linguistic database stores previously translated segments with the express intention of aiding the human translator.
The segments can consist of sentences, paragraphs, headings or titles and the TM is designed as a way of speeding up future translations as all previous translations are accumulated.
It means the same sentence never has to be translated twice. The source text is stored alongside its corresponding translation in a language pair and these are called ‘translation units’.
How does it work?
When the translator opens a segment, the application will look to see if the equivalent source text exists. The results are typically ranked with a score that expresses the percentage of similarity.
An exact match means the segment is identical, while a fuzzy match means the segment is similar, but not identical. If either of these matches are found, they are instantly extracted and placed within the target file. The suggestions can be accepted or overridden, while any segments in the target file that do not have a match are translated manually and automatically added to the TM.
What are the benefits of a TM?
As already mentioned, the whole translation process is much quicker as it removes the need for the unnecessary re-typing of all the existing translations. Knock-on benefits of accelerating project completions are that firms can take on more work and so boost their revenue, while they can offer more competitively priced localizations.
The measure also creates a higher quality final product, as it ensures a consistency of approach within the translations. Linguists using a TM will only be offered translations that have been approved by the customer.