• State Verified Answer
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 433 subscribers
  • Views 4024 views
  • Users 0 members are here
Options
Related

Translation Results window in Editor does not recognise fuzzy matches

Diana Camelia Ciupe

Hi! I am running the latest update of Trados Studio 2017 (SR1 - 14.1.6284.6) (Freelance version) on a Windows 10 platform. Since it installed, most of the fuzzy matches are no longer propagated in Editor - they are no longer recognised in the Translations Results window at the top, although they are shown in the Concordance window. This is quite annoying and it slows down my work a lot.
Furthermore, only random matches are recognised (e.g. a segment with 90% match value is not recognised, while a segment with 76% match value is, but this is does not work as a rule).
I would like to mention that I uploaded the same project into MemoQ 2015, and here the matches work perfectly.
Any piece of advice would be highly appreciated.
Rate translation
×
Suggest better translation
×
Moderator UI
×
  • Get AI Suggestion
Parents Reply Children
  • 0 in reply to Andrei Vybornov

    Dear  


    In general the problem has been fixed, but it's not foolproof. The solution was this:

    Change: LookAhead no longer prefetches segments where it has not found a result in the translation provider. This re-enables use cases where a high fuzzy match follows on to the current segment, or where you add a new term on the current segment that you expect to see in the next segment.

    What this is saying is that if you did have an older fuzzy in the TM then lookahead will prefetch the results, and this means if you just translated something that should provide a better fuzzy, or even a 100%, then it won't appear.  The change will only help where you don't have anything existing in your TM (prior to a few segments ago), so in this case it will not fetch the existing old results, it will check what you just entered instead.

    Perhaps this is the scenario you are facing?

    Regards

    Paul

    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
    Thanks for you prompt reply, Paul. Yes, Studio fetches old matches and fails to see the most recent ones (translated a few segments ago). No matter what was the original intent of Lookahead, the result is somewhat shocking for me as a TM-tool user, because finding matches is its core functionality and I could hardly imagine a TM-tool developer would ever dare to sacrifice part of it in return for faster performance. Who would care about a more powerful engine in a car if you take off the wheels? Most recent matches from the current document are the most valuable. TM is supposed to ensure consistency, but Lookahead does just the opposite: you cannot achieve consistency among segments within the same document because it skips the most recent translations. Of course this is only true if you have a populated TM. When ‘… you don't have anything existing in your TM…’ Lookahead does no harm. But then how realistic is this scenario? All my TMs are populated.

    I am still not sure if Lookahead is a good idea at all because skipping most recent translations seems to be in its nature. But the least SDL should do is disable it by default.

    Thank you again for your help and promptness. Always admire your work :)

    Best regards,
    Andrei