Under Community Review

Trados Studio - enable splitting of merged tracked segments

For Regulatory Labelling, we miss functionality of being able to un-merge tracked segments.

Currently, it is possible to merge tracked segments, but not to split them back to their original segmentation. Linguists sometimes merge segments even though they are instructed not to. This causes significant leverage losses when using "Keep initial segment origin and status" at perfect matching: if linguists merge tracked segments in what later becomes reference file, we cannot un-merge them there, but then the new source segments without a merge for such segments, and we lose leverage on such segments (and also sometimes segments around them, depending on file complexity/repetitions in content).

If we are able to merge, we should be able to split.

Before merge:

Screenshot showing a text file with tracked changes, including a segment that may be incorrectly merged by a linguist.

After merge:

Screenshot displaying a side-by-side comparison of a text file with tracked changes before and after a segment merge.

Leverage loss in new version:

Screenshot illustrating the leverage loss in a new version of a text file due to the inability to unmerge a previously merged segment.

Jurgita Kunsmanaite
Parents

  • I would think the solution is to better prepare the source... What linguists do, is necessary for the target language, at least for m in Polish, where the word order may significantly differ from the source. Confirming segments with no real correlation in meaning is useless for any TM work

Comment

  • I would think the solution is to better prepare the source... What linguists do, is necessary for the target language, at least for m in Polish, where the word order may significantly differ from the source. Confirming segments with no real correlation in meaning is useless for any TM work

Children
  • in reply to Jerzy Czopik

    Hi Jerzy, unfortunately as this is language-specific the team cannot prepare the source in a way which matches the grammar for all 24 languages in EU (or whichever many target languages a particular project handles, but 24 is pretty standard in the team). This means some target segments are split at commas, compared to the source full stops.