Under Community Review

Allow duplicates or create write-protected TUs

As explained thoroughly in this thread https://community.sdl.com/product-groups/translationproductivity/f/90/t/15570 I would like a feature to be added to allow duplicate TUs for source/target pairs that are the exact same. Currently, such feature only exists if the target differs. This feature could be present during alignment imports into TM or within the Update Main Translation Memory batch task. As explained in the thread, keeping the original TUs is essential for maintaining quality and reliability with our translation. All our TUs have robust, detailed fields, therefore "allow multiple values" is impractical. 

Another option could be to "add new translation units if the field values differ"

Or "add new translation unit as write-protected TUs" 

The latter option would prevent important data from being overwritten. I'd like certain documents to always stay associated with a certain project and currently, adding a TU with the same source and target but with different fields from a different project would overwrite the original TU. 

Please consider this addition.

Keenan Cooper

  • Although I understand your problem I do not think this is really useful as it would clutter result window and as Paul said is hard to implement. There might be better solutions to avoid this, so please check my suggestion on your metioned post. Maybe you could set another suggestion with that part in mind. I would support that one as it's more userfriendly.

  • Thanks for posting this idea. While I understand the thread, I will also need to say that this idea conflicts somewhat with the core design of how the Studio TM engine works. "True" duplicates are a "no go" for most (if not virtually all) of our customers. That's why very significant effort went into the TM engine, and how the metadata fields work, over the years to ensure that users get both flexibility to have multiple translations with different targets (very important) while at the same time avoiding duplicates wherever possible, and merging field contents so that where duplicates exist, the metadata can be merged. Also, context matches are available to keep a flow of matching TUs in the same context in a document - metadata fields are not required to have this functionality. Ultimately, these different features work together to try and keep TMs as clean as possible.

    I do think the only best way (for the foreseeable future) is to work with multiple values, and/or try to organise metadata in such a way that you don't end up having a myriad of metadata values. You probably know this, but just to make sure: to convert an existing TM, it's not possible (as you found) to just change the field setup, as this will wipe critical metadata from the TM. Instead, you would need to a) export the TM contents, b) create the new TM to allow multiple values in the corresponding fields, then reimport the TUs. This should give you a TM where the metadata is merged so you can keep track of the values of each TU. This is the fundamental design at this point. Maybe it's worth a try just to see how this may work in your case.

    Another option (which I think you discarded, but just to say) is to work with a read-only TM where you may be able to keep 'true' duplicates as well. But I am not sure that would be practical.

    Thanks again, Daniel