CafeTran also uses pipes for stemming. For example:
perform|ance matches perform, performance,
re|code| matches both recode and code
so perhaps your memoQ glossary is compatible in this regard.
Uh, sorry, my mistake, I should have written "Before using it in CT" instead of "Before using it in memoQ".
I now have the CT glossary with this unintended massive stemming. It looks like this
source term| <tab>target term
And it has hits as described, but I do not want them. What can I do?
Just edit your text glossary in a text editor and search & replace all | characters with nothing to remove them.
Okay, I removed all pipe characters and the problem has been solved. Now the issues have gone and I add pipes only in certain cases (but even then the problem of "too much fuzziness" is not there. I only can assume that – when searching and replacing, see above – there has been introduced a pipe character somewhere in the glossary where it shouldn't be.
I imported my main glossary from memoQ (only the terms, nothing lese).
Before using it in memoQ, I used massive stemming (replace "\t" with ”|\t") as it seems to have more advantages than inconvenience in my case. Then I copied the glossary and inserted it into my CT glossary. I optimized it (remove duplicates, remove source=target), and now I have a kind of fuzziness I do not understand
For "perform" it shows up the term for "form". For "recode" it shows up the tern for "code".
There are no regex terms in this glossary.