Tips

Back in the day, Trados 2007 was the leader in the market of CAT-tools. It consisted of a whole set of powerful programs: TagEditor was used for working with texts, Translator’s Workbench—for working with translation memories, MultiTerm—with terminology, there was a separate module in the form of a Word template and several additional programs. As times go by, Trados 2007 is antique though it is still being used here and there.

In 2009, SDL company launched the first version of a “cat” of the new level—Trados Studio 2009 which broke the neck of Trados 2007. 

It is remarkable that Trados 2007 still remains a proprietary program. SDL neither releases it nor shifts it to the category of open software—probably, that is because it is perfectly functional and able to compete with the major SDL product—Trados Studio.

However, many translators of classic vintage still keep some translation memories in Trados 2007 text format, and sometimes there arises a need to convert them to TMX (translation memory exchange) format for importing into any other “cat” or CAT-system. But how to do it if you got rid of Trados 2007 long ago?

Very easy:

• Start Xbench. Every technical translator is familiar with this tool. What is more, its free version is perfectly enough for the convertation (you can download the installer from here).
• Create a new project in Xbench and upload the Trados 2007 text translation memory:

• Choose Tools > Export Items (or just press CTRL + R):

An Export Items window appears:

• Here fill out all the necessary fields—specify TMX format, languages of the source and target texts, the name and path to the TMX file—and press ОК.

You get a TMX file containing all segments of the old Trados 2007 text translation memory. As long as TMX is a generally accepted international format, this file can be imported into any “cat”.

There are two notes:

1. To perform this kind of transformation, the old translation memory is to be a text one—i.e. of TXT format. Unfortunately, Xbench appears back-strapped when facing a 5-file memory.
2. If you need to apply an old translation memory in Trados Studio, all these operations won’t be necessary. An heir of Trados 2007, Trados Studio supports this format directly. To do this, enter Trados Studio project settings: Project Settings > on the left panel Language Pairs > All Language Pairs > then in the right part of the dialog box press the Add key > File-based Translation Memory...:

A window appears in the bottom right corner; choose the type of Legacy Trados Translation Memory (*.tmw):

State the path to the Trados 2007 5-file translation memory TMW file. Then press Quick Upgrade, wait for all segments to transform, agree to delete temporary files and press OK.

As a result, the old Trados 2007 5-file translation memory is transformed into Trados Studio format. Another file with SDLTM extension will appear in the same folder where files of the translation memory have been stored—this is the very old TM in the new format.

A Wordfast TM file has .txt extension. Actually, it is an usual .txt file delimited with tabs in a special way. Converting it into TMX format that allows TM exchange between all other CAT tools can be done in several ways. The easiest and quickest way involves Xbench (either free or paid version).

Step 1: Load your Wordfast TM into Xbench. There is a special file type Wordfast Memory there:

Do not forget to set the checkbox Ongoing translation.

Step 2: In Xbench, select Tools > Export Items, or simply press CTRL + R.

Export Items dialog box appears. Specify the settings in Output section: what languages are considered to be source and target, and the path and the file name for the .tmx file of the TM.

Press OK button. The .tmx file you get is the converted Wordfast TM.

Recently, we learned how to extract .xlf file from .mqxlz. XLIFF file in memoQ. It appears to be that extracting .mqxliff file from a memoQ delivery package, .mqback file, goes in a similar way.

Like some other memoQ files, .mqback file is actually an ordinary ZIP archive. If you add .zip extension to its name and then unpack it, you get a set of .mqxliff files and a “technical” .xml file (like DeliveryPackageInfo.xml).

.mqxliff files can be processed like any other bilingual files. I.e., you can import them to memoQ project, run QA in Xbench (even in its free version, 2.9), etc.

When your translation performed in Word with Trados (to be exact, Translator's Workbench) is finished, you create a translated file that contains the translation. I. e., you eliminate the source text and leave the translated text (target) only. This operation is called “cleaning”, as you “clean” your file from source, leaving the target only.

Usually, you do it using the Translator's Workbench: you select Tools > Clean Up, then you choose the Word files, press Clean Up. Translator's Workbench deletes the source text and so you get the translated files.

But sometimes something goes wrong and Translator's Workbench fails to clean the files. This usually happens if there are problems with segmentation in the file you are cleaning. Or, you try to clean the files of Word 2007 and higher (.docx files). Or, you need to clean the file from source on a PC where Trados is not installed at all. Then, the following trick helps.

1. Usually, you see only visible text in Word, like this:

You see the translated text only, while the source text is present in the file as well, but it is hidden. To see the source text, you need to make hidden Word text visible: press  button on the Word ribbon or press CTRL+SHIFT+8 (just 8, not F8). You will see something like this:

2. To get rid of the source text is in fact to get rid of the hidden text. This can be done right in Word, without Trados. To do this, press CTRL+H to open Search and Replace dialog window, then press More, then press Format button > Font > Hidden:

Then, press OK and Replace All.

This command deletes all hidden text from a Word file, leaving only visible one — i. e., only your translation.

The easiest way to perform such conversion is to use Xbench. The free 2.9.0. Xbench build is enough. Yes, this tool is good not only for QA :) You do not even need Wordfast itself.

1. Load the .txt file containing the offline Wordfast TM to the Xbench, as if it is a regular bilingual file. (If the TM is big, this process can last for some time.) Xbench recognizes it:

2. Export the loaded segments into the .tmx file. To do this, press CTRL+R or select Tools > Export Items:

In the Language section, choose source and target languages of the TM.

The result of this operation is the .tmx file containing the segments from the initial Wordfast .txt translation memory. TMX is the industry standard format for exchanging translation memories data. Actually, TMX = Translation Memory eXchange. Now you can import it in any CAT tool.