While I have nothing to suggest to you beyond use of the scientific principle (i.e., make the smallest, most elementary application and see if that works) what we are looking at here is something that VFP programmers do in fact run into from time to time ... and it brings to mind a situation which I feel is relevant.

I once found myself in a debate with some big hot-shot C++ professor. I told him there was no such thing as object orientation. He almost didn't reply, but then asked me to justify my statement.

I told him there was no instruction on the Pentium chip set to define or create an "object." Since we were talking about a program that was designed to compile and run on an Intel processor, he had to concede.

You see, if there processor still functions in essentially a linear manner, then any reference to an "object-oriented" language must be made tongue-in-cheek. It is a facade, even in a strongly object oriented language like Smalltalk, then diminishing as the object model becomes less stringent as in C++ or Java, or Visual Foxpro. The so-called object does not really exist; it is a fallacy of either the operating environment or some run-time engine which creates an illusion of object-orientation but actually in the end generates code that will be exectuted in essentially a linear manner. It is a tool for those who do not understand the processor to write programs. That there is so much concern over "business rules" then one might conclude that such things were not designed with the true programmer in mind at all ... it is like a wizard designed so those who cannot code a screen to create a screen, or for those who cannot write SQL to create a query. It is, at best, a crutch designed by marketing so that the software might be sold to those who previously would not have been able to use it.

I think that crutch, at least in VFP, has flaws. Those flaws are concealed with eye shadow and rouge, and there are those of us who stumble upon an area where the makeup is thinly applied.

But what do I know? In the past, I believe that I cured the problem by making variables tighter in scope such that the memory manager had an easier time of keeping track of the particular object hierarchy. Maybe when the VFP "object" accessed public variables it messed things up. Truth is, I really don't know WHY and I hate it when that happens!