• State Suggested Answer
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 433 subscribers
  • Views 4990 views
  • Users 0 members are here
Options
Related

Exact functioning of "Update Main Translation Memories" batch task

Raphaël Toussaint

Since the Studio documentation is as usual very detailed (sarcasm OFF) I was wondering if somebody could enlighten me on the exact functioning of the option "Merge translation units" if target segments differ.

I noticed that in some cases during the TM update, the target of an existing TU gets updated (e.g. when a typo was corrected) and in other cases I end up with "multiple translations" or duplicates in the TM, i.e. having TUs with the same source but different targets, which is wanted when translations differ depending on context.

What defines when an existing TU gets updated and when a new TU is added? I am under the impression that context information is taken into account, but I don't think that it is only that.

The main objective of my question is to have as much control over the TM and its update process as possible to avoid the creation of unnecessary duplicates. In the worst case, I might end up with one TU containing errors in the target and one with the correct target instead of just one which has been updated with the correction…

Thanks a lot in advance for any light that is shed on this mystery!
Rate translation
×
Suggest better translation
×
Moderator UI
×
  • Get AI Suggestion
Parents Reply Children
  • 0 in reply to Evzen Polenka
    I didn’t know either Evzen. I was on a train for a few hours this morning in Romania so spent some time testing all the configurations I could think of. Was a little confused by my results at first until I discovered this.

    Paul Filkin | RWS Group

    ________________________
    Design your own training!
    You've done the courses and still need to go a little further, or still not clear? 
    Tell us what you need in our Community Solutions Hub

  • 0 in reply to Paul
    Thank you for checking this, ! I will try to confirm your findings with my own tests, either today or on Monday.

    In theory, this should be good news for us since we always use the same TM User ID for finalizing and TM update batch tasks and not the information from the SDLXLIFF files. It could, however, result in numerous unwanted duplicates in case a PM accidentally uses a different user ID or the info from the SDLXLIFF files when running said batch task.

    It would be very useful if such crucial information would find its way into the official Trados Studio documentation.
  • 0 in reply to Paul

    It would be definitely good to get this (dis)confirmed by some developer after reading out the exact algorithm from the actual source code.

    The more I think about it, the more silly and short-sighted it sounds... :-\

    Like, while it might sound great from the perspective of updating translator's personal TMs, it is very bad idea in the wider context - when a company/LSP maintains own TMs and sends stuff for translation to 3rd parties.
    Translations can end up being performed by dozens of different people over time and if their User IDs would influence the content of main company's TMs, that would be a BIG MESS :-O
    Especially because the option to use information from bilingual file to update TU system fields is ON by default...

    If one would have to overwrite the User ID before/when doing main TM update - and loose very important ORIGINAL information about the author and time of the actual segment - that would be VERY bad...

    I still can't believe it's true... it would be totally crazy then... :-\
  • 0 in reply to Evzen Polenka
    Absolutely, that's exactly our use case. We decided to use a uniform user ID because too many of our translators have long or cryptic user names (often something like "computername\username", e.g. "LAPTOP-E9QF9SIP\john_doe" or "XY-PC\XY") since they never bothered to customize the one Trados automatically grabs from the system settings.

    A feature I would like to see is that existing TUs in the TM are always updated unless the translator has marked them specifically to be added as duplicate TU during TM update. This would make things way clearer for everybody. I should maybe add this to the suggestion box of SDL, i.e. Translation Productivity Ideas.