Maybe it's me who doesn't understand the way it works but, I'll explain my issue:
My workflow is as follows:
Each time I start a project, I set a ProjectTM (where I store the current project) and a ProjectTerms (for term consistency). I also open a Recall Memory (as read only) where I store all ProjectTMs once everything is proofread and clean to use it for getting matches. As I confirm a segment, the segment is stored in ProjectTM without any problem BUT as the Recall Memory is huge, there's a small delay when travelling between segments, so I do a Preliminary Matching using Recall Memory. Then problem is that when I do that, my ProjectTM stops storing the project segments (actually it stores them but I cannot see them till I close and reopen the ProjectTM). Am I missing something?
Just make sure that you perform the Preliminary Matching only for the larger recalled memory and leave the smaller ProjectTM in the Automatic mode. Then, the new segments added to the ProjectTM will be recognized immediately.
I worked! Thanks Igor. What I don't get is the "reason" of Total Recall, I mean, it's apparent that you can work with larger files and so on but at the same time, you loose some tmx functions such as remove duplicates and tmx clean and even exproting tmx, am I wrong? Is there any problem not using Total Recall and using instead a traditional tmx system ProjectTM/MasterTM?
> What I don't get is the "reason" of Total Recall.
As long as your installed RAM memory is large enough to load all the translation units from the TMX file, you are fine with loading them this way. Total Recall retrieves from the database only the units needed for the current project context so this solution is preferable for the users who have a large base of translation units and wish to avoid loading them all into RAM memory.