Start a new topic

Useless fragments


I please need a little tip about how to avoid this kind of fragments (e.g. "on" and "to" inside a word being listed under fragments with the translation of "on" and "to" found somewhere in the TM) :-).

Thanks in advance!

(13 KB)

Hi Elisabetta,

If you find matches for 'to' and 'on', simply delete those fragments from your translation memory - after right-clicking at the fragment match number. CafeTran shows them because they are there.

Thanks Igor, yes, I'll do that, but is there no way to avoid that such little elements appear as matches as part of another word? For example CT could detect "or" in "word", but this is not very useful ("us" in useful) :-) in my view. Is there anything I can do in Preferences - Memory to avoid this kind of match?


Highlighting of matches in the middle of words is not easy to solve as there are languages without a word separator - without word boundaries. In your case, these are probably exact TM fragments (not some guessed hits whose minimal length can be set in Preferences > Memory). Hence, my tip to remove them from your TM. Another solution would be to let the user set the minimal length of the highlighted word in some future update. 

Thanks for your reply, and the minimal length for "hits" can be set in Preferences > Memory > Subsegment minimal length, right?


That's correct. However, this applies only to the virtual "guessed" hits (not exact fragments such as para=for or en=on present as real translation units in your TM. 

Yes, thanks Igor!


Login to post a comment