>>I suspect that Rick's anticipated performance gain includes allocation of memory external to VFP's housekeeping so you can process quickly with persistent pointers and minimal re-reading of start positions or error check or other overhead,
>>Correct. It's the only way to fly.
>
>Chen agrees with the concept of shelling out to C/C++ before returning results to VFP. He pointed me to this antique link http://www.dfpug.de/konf/konf_1995/oop/SessionE-C%2B%2B.htm ... note reference to the mighty Pentium 90 running Windows NT ;-)

Would like to meet Chen someday.