>Hi Dan,
>
>>If VFP is using a memory structure internally, why does it need to keep the VCX open?
>
>I can imagine a number of reasons. But before I should make clear that this is pure speculation fromt this point on.
>
>The VCX is used to create a class structure in memory. The structure (or maybe a C++ object) in memory is much more efficient to deal with in a language like VC++ in which VFP is written. That's why VFP can't work with the VCX alone.
>
>The VCX could be kept open, because eventually pointers in the structure point to parts of the VCX that is loaded in the cache, eg. the compiled code. Moreover, when you create a class, VFP has to dermine the entire object and class hierarchy. I could imagine that doing so in the tables is an efficient way. Finally, if you could modify the VCX class while it is loaded, the class would go out of sync with the memory copy. Depending on wether VFP uses the internal version or the VCX file when loading a new class, you would see all kind of strange behavior.
>
>Christof

I had heard that they were using a memory map which was pointing back to the VCX, but Ken Levy says that it is primarily for security, which is in agreement with your last comment.