Pages

Dec 25, 2011

Trados TagEditor: Optimal translation of memoQ bilinguals

With the growing number of translation agencies, direct clients and outsourcing translators adopting Kilgray's memoQ as a working platform for managing translation project content, it is particularly important for these new memoQ users and their partners to understand the best approaches to working together with persons who use other tools. One tool which is still commonly found is SDL Trados TagEditor. Compared to the other "classic" Trados tool, the Workbench macros for Microsoft Word, TagEditor has the advantage of enabling many different file formats to be processed while protecting their formatting elements (also known as "tags").

SDL Trados TagEditor can work with two types of "bilingual" files prepared in memoQ: XLIFF (*.xlf) files and bilingual RTF tables. Each approach will be presented here along with some suggestions for best practice.

XLIFF files
TagEditor comes with a default INI file for translating XLIFF, typically found at the path C:\ProgramData\SDL International\Filters\XLIFF.ini.This INI enables the contents of the target segments from the memoQ XLF file to be translated as the source in TagEditor. Thus for this approach to work, the source must be copied completely to the target in memoQ before the bilingual XLIFF is created using the Export bilingual function of the Translations page. This makes pretranslation undesirable in most cases, because the source text for matches will not be accessible and the translator will end up with a very screwy TM. Data for the TM should be supplied to the translator as TMX; be aware that match rates for the segments in TagEditor will differ significantly in some cases.

The memoQ XLIFF files will have a lot of "junk" at the top of the file when viewed in TagEditor:

Skip the content between the mqfilterinformation tags and do not change it in any way. Place the  cursor below that to start working. If you prefer not to see that information at all, use the XLIFF INI for TagEditor which I modified for use with memoQ XLF files. Then the XLIFF will look a bit cleaner with the header information filtered out:

Astute observers may have noticed, however, that all is not really well with the tag structures in the views above. I think there is  problem with the way that memoQ is generating the XLIFF files, with some tag structures being replaced by entities. (You see this if you open the XLIFF from memoQ in a text editor.) This causes consistent problems like the following in TagEditor:


This will require a lot of tag fixing. Thus I really can't recommend the XLIFF method at this point, not for my simple little test file in any case. The methods using the bilingual RTF tables with memoQ tag protection are safer and the structures that result are much simpler.

But if you do use this method, when the translation is complete, clean the TTX file using Trados Workbench or use the menu option File > Save Target As... in TagEditor to create an XLIFF file to return with the translated content. If the content inside the mqfilterinformation tags has not been segmented, an accurate count of the words translated will be shown in Trados Workbench upon cleaning the TTX (as accurate as that tool is given its limitations with numbers, dates, etc.)

Bilingual RTF tables
There are created in memoQ using the Two-column RTF option of the Export bilingual function. Technically speaking, the files have more than two columns (source and target, index numbers and possibly columns for a second target text, comments and status). Good practice for working with these files in TagEditor and many other tools also requires the source to be copied to the target column. This can be done in memoQ or later in a word processor. The table might look like this, for example:


For best results in TagEditor, it is important that this file be generated with the "mqInternal" style selected for tag formatting. The dark red color imparted to the tags with this option means that proofreading in a word processor is easier, and it also enables the text of the tags to be selected and hidden using a search and replace function. If the RTF file is then saved as a Microsoft Word file, the memoQ tags in the table will then be protected in TagEditor!


If the "full text" option for tags is selected, this makes little or no difference in the TagEditor view.

Here's a quick look at what the protected memoQ tags look like in TagEditor and what can happen without protection:




One possible workflow for memoQ RTF tables in SDL Trados TagEditor consists of the following steps:
  1. Copy the source text to the target in memoQ
  2. Export a bilingual "two-column" RTF file with the mqInternal style option selected for the tags
  3. Re-save the RTF as a DOC or DOCX file! This is necessary so that TagEditor will use the right filter.
  4. Select and hide all the text in the file
  5. Select only the text to translate in the target column and unhide it
  6. Using search and replace, hide all the dark red text. The settings for the dialog are show below and are set using the Font... option (marked with a red arrow in the screenshot) in the Format dropdown menu of the Replace dialog.


    The font color to hide will be found under More Colors... in the font colors of the font properties dialog:

  7. Launch TagEditor and open the Microsoft Word file with your content to translate. All the hidden text will be protected in tags. Translate the accessible text.
  8. Create a target MS Word file from your TTX as described above for the XLIFF files translated in TagEditor.
  9. Open the target file and unhide all the text.
  10. (Optional) When reviewing the text in the word processor, comments may be added if there is a comments column. These will be imported back into memoQ and can serve as valuable feedback.
  11. Re-save the target file as an RTF
  12. Re-import the RTF with the translated table into memoQ. The target text will be updated to include the translation. 
  13. A QA check for tags, terminology, etc. should be performed in memoQ before exporting the final file for delivery. If an external reviewerr is used, another bilingual file in an appropriate format can be generated in memoQ for that work.
Steps 4 to 6 can be performed using a macro for convenience.

The procedure described above can, of course, be abbreviated considerably by simply copying the source text cells into a new Microsoft Word document, doing the search and replace to hide the dark red text for the tags, then processing the file in TagEditor. After translating, unhide the text in your working file, then paste the cells over the target cells in the RTF file.

Here's a look at the test file translated in TagEditor (with a comment added as shown by the dark speech balloon icon) after it was re-imported to memoQ:



And here's the translated file itself:




4 comments:

  1. Great post and thanks so much for sharing. I believe Trados should support MemoQ project files in Trados Studio. I see that this can also be a good project for an exchange plug-in.

    ReplyDelete
  2. Well, in a sense, Studio already does - no plug-in needed to handle the memoQ XLIFF bilinguals. It would help, though, if the bug in Studio were fixed that causes difficulties reading an XLIFF without specifying the sublanguages. That has been ignored by SDL far too long; when I spoke to one of the developers in Warsaw last September it was clear that this compatibility issue wasn't a priority. But of course a project consists of more than that. I hope that both memoQ and SDL Trados studio will support the Translation Interoperability Protocol (TIP) for project data (translatable content, TM and terminology) exchange some day.

    ReplyDelete
  3. Thank you soooo much! I tried by setting the source column and the ID column to DO_NOT_TRANSLATE style, then I tried setting them to Trados' tw4winExternal style, but in TagEditor the result was a mess. Your hint with "hidden" works far better.


    Greatly appreciated!

    ReplyDelete
  4. Update: actually, it works with tw4winExternal style as well, but the your trick save as doc is priceless!

    ReplyDelete

Notice to spammers: your locations are being traced and fed to the recreational target list for my new line of chemical weapon drones :-)