Translation Memory
Translation Memory
A translation memory is a database consisting of translation units, which are most often sentences in the source language together with their translations into the target language. Before translation, each new sentence is compared against the contents of the translation memory (i.e. all previously translated sentences) in search of similarities.
Upon translation, each source sentence is written to the translation memory together with its translation. A translation memory hosted on a server enables multiple translators and proofreaders to work together on the same project. That allows each participant in the translation process to see not only their own translations, but those of others as well. That ensures high translation quality thanks to consistency of style and terminology throughout, speeds up translation, and cuts costs – the same sentence never needs to be translated twice, and many rounds of proofreading are eliminated.
For over 20 years, we’ve been using translation memories that contain all the translations done for a given client, in a given field, or for a given product. That means our clients’ translation memories sometimes exceed several million segments (a dozen or so million words), which directly translates into reduced costs and turnaround times for subsequent translations, while maintaining high quality.
LocStar uses the translation memory systems of many software companies, including SDL Trados Studio, SDL Trados GroupShare, memoQ, IBM Translation Manager, XTM Cloud and Memsource, as well as LocTM, its own proprietary server translation memory solution.