I have partitioned my data into several files, with these characteristics:

1. One master file containing only direct ancestors. I use this to produce an Ahnentafel chart.
2. A separate family file for each unique family. I presently have about 50 of these.
• All descendants of a single ancestor
• Only of that surname (or variant)
• With intermarriage between families, both partners are duplicated in each relevant family file. 
• Children of intermarriage are represented only in the father's family file.
• Each family file uses a unique 4 character reference number prefix.

Why have I done this?

1. With FTM's susceptibility to file damage, I have limited the extent of damage to a single file. Work can proceed on other fronts while the damage is repaired.
2. Performance. Many functions, such as LOAD, SAVE, operate more quickly with a smaller file.