What is a Translation Memory?
A translation memory (TM) is a linguistic database that continually captures your translations as you work for future use.
All previously translated segments are accumulated within the translation memory (in source and target language pairs called translation units) and reused so that you never have to translate the same sentence twice. The more you build up your translation memory, the faster you can translate subsequent translations and your new project is more consistent, enabling you to take on more projects and increase your revenue.
Translation memory managers, usually included in Computer-Aided Translation (CAT) tools, are most suitable for translating technical documentation and documents containing specialized vocabularies. Their benefits include:
Consistence – Ensuring that the translated documents are consistent, including common definitions, phrasings and terminology. For example, if the software is already translated and now it is time to translate or update its manual or help, the translation memory will use the previous stored sentences to guide the current translation. In other words, the documentation or the help file will be consistent with the terms used in the main software.
Flexibility – Enabling translators to translate documents in a wide variety of formats without having to own the software typically required to process these formats.
Speed – Accelerating the overall translation process; since translation memories “remember” previously translated material, translators only have to translate it once.
Cost – Reducing costs of translation projects. For example, your company just updated the radiology image equipment. Usually, most of the user guide document will stay unchanged, only some items in the document will need to be updated. The translation memory has the ability to use the previous sentences stored in its database to pre-translate the current updated manual. This benefit will cause a relevant impact in the translation cost.
How does translation memory software differ from machine translation?
Machine translation automatically translates a document without any human input (or assistance).
These kinds of tools are fast, but result in a poor quality translation as a machine cannot understand the subtleties or contexts of language. As a result, quality and accuracy tend to be around 50% – 70%, therefore it is not advisable to send the raw form directly to your customers.
For example, the simple text below was translated using Google Translate (one of the most popular Machine Translation tools). The source text was translated from English into Korean. Using the same MT system, the translated sentence was translated back to English (which we call as reverse translation).
Original English text:
“Excel Translations’ unique organizational structure allows the company to handle all phases of a medical translation project.”
“엑셀 번역 ‘독특한 조직 구조는 회사가 의료 번역 프로젝트의 모든 단계를 처리할 수 있습니다.”
Reverse English Translation:
“Excel Translations’ unique organizational structure, medical translation project, the company can handle all phases of.”
However, the “imperfect but fast” machine translation system is efficient in some cases. IBM and Microsoft have been using machine translation engines for the last ten years to provide real-time feedback for their customers. For example, a user can ask a question on the IBM portal about the installation for an acquired product in China. This question (written in Chinese) is automatically translated into English, so IBM’s server’s search engine finds the highest match topics. Then, these topics are automatically translated into Chinese and the customer can find a solution for his question in a few seconds.
In this case, the quality of the translation is poor, but the customer is a 100% satisfied because he is now able to install his product. However, on the other hand, the Machine Translation’s quality is not acceptable for providing a multilingual prospect for a Neurovascular Array system.