What is Translation Memory?

Transifex TM What is Translation Memory Reasons 3 Everyone Should Use it The set up Let's say your translator can translate 2-3k words a day. If your content isn't literature, but rather marketing communications, documentation, a travel listing and so on, then chances are that they will be translating similar phrases multiple times. If the translation of all similar con- tent could be automated or at least referenced, a transla- tor could work much faster and more consistently, thereby saving you money. Sounds like a cool plan, right? Cue the lights for Translation Memory (TM) TM is a technology that was designed with speed in mind. So how does Translation Memory (TM) work and why is it a tool you should be using? Read on to find the answers to those questions! What is it TM is a database of previously translated segments, which can be sentences, paragraphs or sentence-like units (headings, titles or elements in a list). Translators are usually able to see this within their translation tool. This helps them when they are working on content similar to what has already been translated. USED Definitions Your TM compares untranslated segments with previous translations, analyzing it for one of two kinds of matches. Fuzzy Match 100% Perfect Match Sometimes content analyzed When a segment being translated matches a segment already in the TM word-for-word, this is considered a 100% or perfect against your TM is not a perfect match, but it might be close! You might find content that only varies from a previous segment in your TM by 1 or 2 words. This is known as a fuzzy match. match. These 100% matches reduce your overall cost of translating your content (more on this topic later). You will typically get a fuzzy match if 50% to 99% of the new segment matches something in the TM. 1 Translate faster Improve consistency As more content is translated, consistency becomes ever more important. By referencing or reusing previous translations, translators can ensure that With TM, your translators don't have to translate new content from scratch, which means they can trans- late more in less time. How much faster they can translate depends on the number of fuzzy, repeated and perfect matches. Here is an example: new translations blend in with past work, or even work done by another translator. AVERAGE NEW WORDS/HOUR TRANSLATED BY A TRANSLATOR 250 350 Normal Using fuzzy matches Lower costs TM reduces the amount of new work a translator has to do. In turn, you would pay a lower price for segments with a TM match. The standard industry pricing breakdown is: New words - 100% Fuzzy words - 50% 100% matches (a.k.a. Perfects) - 25% As an example, a word that normally costs $0.16 would cost the following: NEW FUZZY 100% ¢10 ¢5 ¢1 ¢1 ¢1 ¢5 ¢1 $1 ¢1 ¢1 ¢1 Ф0.16 ¢0.08 ¢0.04 TMX files First released in 1997, the Translation Memory eXchange (TMX) file format is a standard way of exchanging TM data. So why are they important? Using TMX files, you can move your TM between tools or translation vendors without worrying about losing important data. Sources 1. http://en.wikipedia.org/wiki/Translation_memory 2. http://www.tcworld.info/e-magazine/translation-and-localization/article/ten-good-reasons-for-using-a-translation-memory/ 3. Alicia Deadrick, Business Development Manager, e2f 4. http://ginstrom.com/translation/translation_memory.php 5. http://www.languagescientific.com/language-services-blog/fuzzy-matching-makes-translation-cents.html 6. https://en.wikipedia.org/wiki/Translation_Memory_eXchange Transifex TM www.transifex.com II II