Trados Alignment for 2017 - why is it well-aligned in 1 machine and poorly-aligned in another?

Hi all,

We have Trados 2017 installed on separate laptops. Both installations have different SR1 versions, and we are trying to align the same English and Traditional Chinese document pair (Align Single File-pair). 

The following is the output of Trados 2017 SR1 14.1.10017.54688:

Screenshot of Trados 2017 SR1 14.1.10017.54688 alignment output showing clean alignment with few segment reconnects required.

The following is the output using Trados 2017 SR1 14.1.10018.54792: 

Screenshot of Trados 2017 SR1 14.1.10018.54792 alignment output showing poor alignment with many segment reconnects required.

The first result is very, very clean, with only very few segment reconnects required. The second one is a very poor result, requiring us to take significantly more time to clean up (using Google Translate to help, etc ....). Interestingly, the segmentation in the source and target languages is the same as you can see, it is just that alignment's guessing logic is showing very different results? 

I'm wondering if anyone knows whats going on here? We applied the latest update to the second laptop but didnt expect alignment to behave weirdly. We also have not used alignment on the said machine before the update to 14.1.10018.54792, so we can't tell if it has to do with the Trados installation or something else.

Appreciate if anyone can share a similar experience and the solution please? Thanks in advance!!



Generated Image Alt-Text
[edited by: Trados AI at 12:16 PM (GMT 0) on 29 Feb 2024]
emoji
Parents Reply Children
  • Thank you so much Paul for your inputs! Yes that's best solution/workaround for now.

    I wanted to ask - have you any idea why does Alignment and the standard project analysis seem to work differently?

    What I mean to say is that when creating the alignment, the source segments are relatively clean (very few tags). And then when I create a new project analysis and use the same source which was used to perform the alignment, Trados inserts several extra tags in the source segments.

    This unexpected behaviour makes the leverages a lot less as I was surprised - its almost as if Trados project analysis and alignment follow different segmentation rules? 

     

    emoji