A single project can a maximum of 24 target languages. This is due to a limit in the number of fields a single table in a database can have (due to the Microsoft Jet Database Engine that Déjà Vu uses).
To be specific, the maximum number of fields in a table is 255, and the Pairs table in a DVPRJ uses 31 fields for the source language and various pieces of information, and 9 fields for each target language (for a total of 247 fields).
It is possible to work around this limitation and work with more languages than that in a single project. Though we usually do not recommend this, it is possible to do the following:
- Create the project, specifying up to 24 of the needed target languages.
- Import the files and fix any possible segmentation errors (using join/split).
- Make a copy of the DVPRJ file.
- In the copy of the project, remove the existing target languages and add the ones that couldn't be added to the original project due to this limitation.
- Process the projects as usual.
The idea is that by making a copy of the DVPRJ file you will not only not have to reimport the files (and fix any possible segmentation issues), but the project, file and row IDs will be the same, so when the two projects get sent to the TM, their segments will all be aligned.
If the 2 projects don't have completely different target languages, or they have different files or different segmentation, this can result in segments in the TM being overwritten by different ones, and Déjà Vu finding incorrect matches. Therefore, the instructions must be followed carefully if this is done at all.
Comments