Translation memories (TMX)

The TMX (Translation Memory eXchange) format is widely used in the translation and localization industry for storing and exchanging translation memory content. memoQ can export the contents of translation memories into TMX files. It can also import TMX files into translation memories, and what is more, it can also import TMX files into term bases. A TMX file saved from memoQ can be read in other computer-assisted translation (CAT) programs, and, through TMX, memoQ can also load translation memory files from foreign programs.

The TMX Standard

TMX files are XML (eXtensible Markup Language) files, and they are not normally meant to be edited manually, although the human eye can read them. The format was defined by the Localization Industry Standards Association (LISA: http://www.lisa.org), and the details on the standard can be read here: http://www.lisa.org/tmx. For more information on the XML (eXtensible Markup Language) standard as such, turn to the World Wide Web Consortium: http://www.w3c.org/xml.

TMX has several versions. memoQ can read all kinds of TMX files from version 1.1 through 1.4b, and saves TMX version 1.4b files.

What TMX Files Look Like

A TMX file holds the contents of a translation memory. It consists of a number of translation units (TU). A translation unit contains the text of one segment in one source language, and one or more translations of this text in one or more target languages.

Within a translation unit, each segment (i.e. the source segment and the target segment) is tagged with a language code. TMX uses standard two-letter ISO language codes, with occasional sublanguage (dialect) tags appended. For example, en-us means US English. The complete list of such language codes can be found here: http://www.unicode.org/onlinedat/languages.html.

Note: When memoQ loads TMX files, it must select two languages from the TMX file (in case its translation units have more than two languages). Using the Translation memory TMX import settings dialog, you can manually match one of the TMX languages to your current source language, and another to your current target language. Thus you can also import a German-French TMX file into a French-German translation memory. In other words, you can reverse translation memories in memoQ.

Formatting

memoQ stores formatting with the text of the segments in the TMX files. It saves formatting information into exported TMX files, and it is also able to load formatting data when importing translation memories from TMX files.

SDL Trados™ TMX files

Trados uses a special flavor of TMX, especially when it comes to formatting. In TMX files exported from Trados, translation units can include two special types of formatting:

· Inline tags: Segment text can contain XML-like tags, for which Trados has a special representation so that the Trados TagEditor program can display them properly, and segments in TTX files can be accurately matched against Trados translation memories.
· Rich Text formatting: Segment text can contain Rich Text codes that determine how the Trados TagEditor displays the text and the tags. Trados retains these formatting codes both in TTX and TMX files for efficient translation memory matching, while memoQ does not need these.
 

memoQ is capable of fine-tuning the import of Trados TMX files – in a way that, whenever possible, text imported from TTX files yield 100% matches against translation memories that have been imported from Trados TMX files.

Translation memories (TMX)