Translation Memory

Translation memory (TM) is a tool that allows us to increase efficiency by storing previously translated phrases and terms in a special database for translators to reuse in new translations.

At Argos Multilingual, our TM databases allow us to maintain a consistent standard of quality throughout a translation project, which saves time and lowers costs. TM tools also allow us to estimate the size of a translation and the number of internal repetitions. Based on that, we can better plan the translation project framework.

Depending on the type of project, we will choose from the following TM tools:

• Across
• XTM
• SDLX
• Trados
• DejaVu
• Transit

We utilize translation memory servers to allow multiple translators to work on a single large assignment or portion of a translation project. With TM servers, translators can complete large portions of a translation project simultaneously, speeding up the translation process while at the same time maintaining accuracy and improving the consistency of the translation. In addition, we keep a unique TM database for each client, which we use only for that individual client’s projects. The greater the number of projects we complete for a client, the larger the database, which in turn leads to greater efficiency and lower costs.

TM Consultancy

If you are interested in implementing a translation memory system, we have a range of consulting services designed to make it easy. We can carry out an on-site needs analysis to determine the best solutions for you and then work with you to draw up an implementation schedule. We handle everything, from guiding your choice of hardware and software to training your translators and engineers in the use of the tools, including:

• Training in the use of SDLX and other SDL applications (HelpQA, HtmlQA, SDLTermbase, Toolproof, SDLInsight)
• Localization engineering services

The complexity of the system you need will depend on the number of translators you employ and the type and quantity of the material you translate, as some materials are better suited to TM usage than others.

Creating a Translation Memory Database

1. Our client has an initial version of a software application in English.
2. We translate the software into Polish. Our translators create a translation memory database for this translation, consisting of terms in English with their Polish equivalents. Note that we can also create a TM database based on documents translated previously by other translation companies, a process known as alignment.
3. A new release of the software in English becomes available (version 2) that must also be translated into Polish. We are asked to translate the update, which is easy because many terms and sentences from the first version of the software will probably be repeated in version 2.
4. Having a translation memory database based on version 1 of the software, our TM specialist uses this database to analyze version 2 to find all the terms that were already translated and are repeated. The TM database is applied to the new version and the translator now has to translate only the outstanding text. The entire text is then of course edited and proofread by our linguists. The client does not pay full price for any phrases that were previously translated.