Good evening / afternoon / morning,

I hope this is the correct group for posting this question.

I have a project where I have both en-GB and de-DE segments in the source, and the entire (very large) document must be translated to nl-NL. The segments are all interspersed. Creating two separate documents would be a major hassle.

I am all aware of AnyTM and I have actually been using it since it was only available from CodingBreeze directly, for a fee. I have set up the TMs that I want to use, which one to update and which ones not. Then, in the editor, I used the View menu to select ‘Mixed source language translation’ and indicated that the source contains both en-GB and de-DE. When setting up the project, I selected the pair en-GB > nl-NL. I also created a dummy project in order to set up a de-DE > nl-NL TM, and added that through AnyTM to the project. However, if I set that one also to be updated, ‘Update Main Translation Memories’ throws an error with the message: "TM update batch tasks are currently not supported in mixed-source language projects." But I don't know how to update it as an individual task, instead of a batch task. The help files don't give the answer that I need.

Concordance works well and also checks any of the AnyTM pairs. But, although I have a trilingual termbase, Term Recognition does not recognise any German source words. When adding a new word through Ctrl-F2, the source word is put in the field for English, even if it comes from a German segment, as properly recognised by Concordance. Obviously, I can manually delete the word from the English box, replace it with the proper English word and stick the German word in the German box. But, because it is now in the German box and the same word comes up again later on, it is not recognised as Term Recognition only seems to search the English boxes. Meanwhile, I understand that there is no real way around this, apart from entering each term under all three languages, with an appropriate language marker.

The other problem is the TM itself. All source segments are marked as en-GB, even if they are de-DE, although they are properly recognised as de-DE in the editor window (this is confirmed by the lowest bar in the ‘Segment source language’ box, below in the ‘Mixed source language translation’ window). I have read earlier responses to similar questions, but not found the answer that I need. My objective is ending up with two TMs: one en-GB > nl-NL and one de-DE > nl-NL, without German segments in the English file and without English segments in the German file. Yes, I could simply create a copy of the en-GB > nl-NL TM and then cut and paste the German segments, but there are 654 segments ...

Best regards,