Bill,

Strings of length = 6 are too smal to see the optimization Steve was talking about. You have to get up to strings longer than a couple of thousand bytes to see much performance difference.

I'd do the whole thing with a C++ library call that would beat the snot out of VFP in this case.

>I've run a number of tests now and agree with your conclusion.

lcresult = SUBSTR(lcfoo, lni, 1) + lcresult

is faster than

lcresult = lcresult + SUBSTR(lcfoo, lni, 1)

The iteration structure doesn't seem to have an affect.
>In any case, they are both FAST!