>I hope Garrett Fitzgerald won't mind I quote his blog here:
>http://blog.donnael.com/2006/02/dangers-of-not-using-naming.html
>
>I don't agree that applying a naming convention is the proper solution to the issue reported in the blog. The issue should be addressed by applying the "m." prefix.
>
>Garrett, a VFP expert, spent 15 minutes in finding the problem. Imagine an average (or below) VFP programmer gets unexpected problem when he runs the VFP RI Builder. He may never know the problem is caused by a name conflict between his table field "TcNewValue" and the variable "tcNewValue" of the VFP RI Builder.
>
>My opinion:
> When we are coding for our own project, "m." is not necessary.
> If the code is going to work with others' (clients or public users) tables or codes, the "m." prefix is necessary.
>
>I hope the xBASE (VFP) code of Sedna won't have similar traps (if we don't call it bugs).

How about, for consistency, you apply the mdot in both cases? It's faster and it's safer.