Silvia Barra wrote:
I was wondering whether it is better to keep these large TMs or to split them into sub-TMs (e.g., patents related to medicine, patents related to chemistry, etc.).
I would, yes.
- Just in case - are there some methods to split big TMs into smaller ones, for example, keeping some segments (general) in each sub-TM?
No, you'll just have to create them as you go along.
If you want to split them, you can do word searches to extract likely TUs.
Another tactic may be to simply add "!!! " to the start of every source and target segment, and duplicate the big TM into each subfield, then continue using the entire TM for each subfield, so that you can easily see when a match comes from the "old" section of the TM.
I know that big TM allow me to find some more concordances in my texts...
Your CAT tool should be able to do concordance searches in more than just the active TM. I know WF has this feature. OmT doesn't have it but with OmT you can use multiple TMs and simply give your big TM a warningly name (because in OmT you can see which TM a match is proposed from). I'm not sure about other CAT tools... but I'm sure it must be possible.
...but also they are more difficult to manage and I fear that mixing arguments sometimes not so close to each other can be somehow misleading.
This is a very good reason not to mix different fields in one TM.