CafeTran's default settings do not assume the user will create definitions or notes. It hides this complexity on purpose. The same is true for your glossary/terms memory/segments memory/ organization. You have the variety of choices progressing from the simples text (tab delimited) glossaries to the most complex SQL database solution. I would recommend the following:
1. Use the tab delimited glossaries for simple source=target entries, although more fields can be defined here too.
2. Use TMX memory for terms if you wish to add additional fields such as notes, subject and reference/definition fields.
3. Use the real database for huge term bases (with all the required fields) if you wish to hold several million entries there.
The above division is arbitrary since, for example, even a simple text-based glossary with a million entries should be handled efficiently by CafeTran.
The additional fields can be set in Edit | Options | Database tab. When any of those fields is set, CT will also add it to the user's interface.
Adding entries to the glossary/TMX memory for terms or database cannot be simpler. Just select a source term in the source segment window, target term in the target segment window and press the appropriate button (or a shortcut). CT has also an option to skip the New term window (Translation | Options | Skip new term window). I think only a voice command system might be quicker.
The issue in your case is adapting existing glossaries from another CAT tool. Here, I would go for a tab delimited solution and prepare a .csv file for loading it into the program. The addition fields should also be displayed in the CT interface (the glossary tab) "on the fly".
