|Article Title: Machine Translation vs. Translation Memory|
0, Average Rating: 0 (10 Max)
The computer revolution has quickly added tools to most industries. Linguists have not been immune to this phenomenon. There are a variety of tools and options available, but machine translation and translation memory, are the ones most often confused.
With a plethora of translation tools available, to those outside the translation industry, one program is basically the same as another. However there is a big difference between tools such as MT (Machine Translation) and TM (Translation Memory). One is a computer-generated translation, and the other is a tool used by a professional translator.
Machine translation is freely available at many online venues. The results can vary greatly from somewhat legible to complete nonsense that is more comical than useful. It's not wise to trust important documents to the machine translation engines currently available. Professional translators do not rely on this type of technology. In fact, editing content translated using machine translation would slow most translators down considerably. Is machine translation a bad thing? Not at all, it has it uses, but it won’t be replacing human translators any time soon.
Translation memory is based on the concept that documents can contain identical or near identical sentences. In simple terms, the program keeps track of the translation entered by the translator for each sentence. If an identical sentence is later found, the program inserts the translation entered by the translator, who can either use the sentence as is, or edit it. The program can also distinguish if a sentence is 80 or 90 percent similar to a previously translated sentence, and suggest what is known as a “fuzzy” match, which the translator can choose to insert and edit to fit the new sentence. Translation memory is basically a database, which can contain translations from more than one document.
This technology has it’s advantages. With some documents it can increase translation speed by as much as 30%. Translation memory works best with documents that contain a lot of repeated phrases, such as certain technical manuals. These tools generally have terminology managers built in, allowing for access to a custom glossary while translating, all within the one interface. However, there are disadvantages and things to keep in mind when using translation memory. It tends to be more difficult to obtain an overall view of a document translated in this fashion, making it possible to end up with a document that doesn’t flow properly. Anytime translation memory is used it’s advisable to carefully review the document, beyond the normal proofreading process, in order to make sure it flows smoothly.
I hope this has provided you with a better overview of your translation options, and some of the tools available to professional translators.