Translating software: automatic translation and translation memory
Software translation is a translation of an application's text resources. Generally, translation is an integral part of the localization process. Localization, however, is customizing an application for a specific locale, for example, Continental Chinese or Chinese spoken in Taiwan . Besides, localization includes customizing user interface, colors, images, time, date and currency conventions, so that the localized application is fully-attuned to the target consumer group.
Translation of an application's user interface, documentation, online support is sometimes an expensive and time-consuming process. In order to speed up this process, the translator may use specialized tools, such as Machine Translation (MT) or Translation memory (TM).
Machine translation, also called translation software, automatically translates the text content into the target language. Machine translation takes into account the grammatical structure of each language and implements specific rules as to convey the grammar of the source language into the target language. Machine translation produces so-called draft-quality translations: the resulting text can be readily understood, but needs a considerable amount of proof-reading. Thus MT is not a proper solution for the translation of literature or legal texts. However, some MT software tools give you up to 80% accuracy. In any case, the proof-reading of the automatically translated text should be done by a human translator.
Machine translation differs from word by word translation, which translates each word (or phrase) of the source text separately, and does not consider the grammar structure of the source and target languages. Word by word translation can be of a considerable help for the translator, though it is not an automated process.
Computer Aided Translation (CAT) is sometimes called interactive translation and often consists of terminology management and translation memory to enhance efficiency and accuracy of translation, especially in longer documents.
Translation memory consists of a database that stores source and target language pairs of text segments. This database stores words and phrases that have already been translated, in order to re-use them in a future translation. Translation memory is a highly effective solution for the types of texts that are to be reviewed or updated in the future. Though the standard unit of text in a TM database is a sentence, the translation memory user can define what the unit will be. After a translation has been performed, TM creates a glossary of terms and individual sentences stored for future recall. Terminology database, included in the translation memory, allows the translator to keep the terminology consistent throughout the entire document, which is very important for large translation jobs, with a number of translators working over the project. Without a terminology database compatible with the translation memory, the TM user will probably need a separate electronic dictionary or a conventional dictionary.
Translation memory appears to be a very time and cost-effective tool when used in conjunction with specialized localization software tools. Translation Memory solution, applied in Lingobit Localizer, allows the user to leverage his or her work on previous translations, which makes localizing minor updates of the project close to automatic process.
In order to allow easier exchange of translation memory between tools and translation vendors without any loss of critical data during the process, most of the localization suites, including Lingobit Localizer, use TMX standard. TMX ( T ranslation M emory e X change) is the vendor-neutral open XML standard for the exchange of Translation Memory data.