>>>Here's something interesting, though - for one-off handy-dandy programs I was using VFP9 rather than C# because - I thought- of its relative simplicity and brevity.
>>>In fact, now I've been programming C# 99% of the time I've now begun using it for the one-off's.
>>>I was just more familiar with VFP, it seems.
>
>Tis true in many industries that you offer what you've got, you extol the virtues of what you know. But in our case, after clumsy initial NET data was replaced by typed datasets followed by L2SQL that got Foxed, we weren't prepared to put more eggs in the MS basket. Forking in open source can be just as shattering and I exhausted myself in Java in the early days when releases were relentless and IE updates would break Java. Whereas VFP? In VFP I have a prg called pread6.prg that was written in 1991 and has been added to ever since, with original routines surviving largely intact. That's worth a lot IMHO. Especially now that VFP Compiler lets me compile it to C++ that seems to run quicker than the already blazingly quick Fox.
>
Chen (when I last looked) did not claim large speed benefits - which I concluded stem from still using NTI for memory variables and the vfp stack always having to decipher those parameters. C code in .fll is not markedly faster if you access you local variable in vfp-compatible memory slots, the speedup does occur if you use C data types and C arrays (memory areas with pointers).

Where do you see speed up via Chen compiler vs. VFP runtime?