>As I understand it the limitations encountered are related to the general architecture of VFP6 and that is why they won't be changed in this version. For example, the internals of VFP would need to be greatly modified to change the way VFP handles arrays in order to allow safearrays.

Jim,
I agree that the improvements are worthwhile. I'm something of a VFP bigot, I suppose, because I feel VFP is the the best business application development tool around (I don't write compilers or embedded apps, for example), but the safearry issue galls me only because I am presently interfacing with other languages instead for doing purely VFP work. By the way, I would be really suprised if VFP doesn't use safearrys internally since VFP is written in C++, VFP arrays "act" just like C++ safearrays and VFP functions will accept safearrays perfectly well.

I am resigned to incomplete COM support because most VFP developers don't work in this area intensively, if at all, and don't clamor for this support. I certainly didn't until my current client needed this functionality.

Here's to continued evolution of a great tool.

Gary Foster
Pointsource Consulting Inc.
gary_foster@starkey.com