This is something that I struggled with for a while.
I have decided against using George's system for the GFA files for the following reasons.
1) It's a pain in the butt to track backward for several generations. Since we tend to do this quite a bit, it's a significant issue.
2) When I started the database, I used the Case numbers, and redoing all that (or just adding Farrell codes) just doesn't float my boat.

3) The main reason is this:
There are many people missing in the Case data and we hope to fill many of those holes. With Farrell codes, if we find another child of a couple, all the codes from that point forward, for that line, (and all the younger siblings of the Goodrich parent of that child), have to be changed since it is based on birth order. I just don't see it as being feasible. I am going to use a variation of the Case numbering system, something that has the flexibility needed to handle those types of issues.
Don't worry about trying to emulate our numbering system. I am in an unusual situation here. Use whatever you like Richard. I suggest you continue to use Farrell codes. If you send Delores or I data with that numbering system (or any other system for that matter) we will be able to use it without issues.

Matt