Walter,

*deep sigh*
If you admit that you didn't do any benchmarks, how can you be so sure? ;)

If you can find the specific threads that reference benchmarks, I'd be more than happy to read them and try them out.

All I can state with certainty is that I've done some pretty extensive data manipulation in ADO.NET and VFP, some in production and some in tests, and really have not found any significant differences.

If there are cases where performance differences by a factor of 2, I'd truly would like to know, for my own benefit, should I ever have a situation in production where it would be relevant.

'the solution consists of more lines....is likely to be slower as each line has to be interpreted...in VFP a comand is interpreted once'

This is one of the most common arguments I hear regarding the replication of VFP functionality in another tool...and more often than not, it's a facile argument promoted by supporters of the original product. It's also one that is seldom stated conversely.

Keep in mind that there is a bit of overhead involved in these 'xBase' commands, because VFP does some data-checking at runtime. I can do a LOCATE FOR InvoiceDate = "KEVIN" in VFP, and it won't get checked until execution. Yes, it's a very quick check...but it's something that's done more at compile time in C#. There could be many 'efficiencies' and 'inefficiencies' under the hood...do you actually HAVE the source code for the VFP LOCATE? ;)

Have you ever had a need to write a 10-20 line function to do something in VFP that existed as one line in another language? I've had to a few times, and would have found the arguments to be equally weak.

The bottom line is that I can do a LOCATE with 50,000 rows in one line of code in VFP and roughly 10 lines of code in C#/ADO.NET, and get the same results. Once again, much of what's under the hood of ADO.NET traces back to VFP cursors.

Walter, I'd feel much more comfortable continuing this discussion if I felt confident you were trying some of these first-hand.

Please remember I am not saying that data munging in ADO.NET is always as fast as VFP (or vice-versa!). All I can say is that my research has not found noticeable differences. But this thread is here for us to share experiences and learn from them...and I'm truly interested in the experiences of others.

In the last ten years, did you ever have someone who worked in another tool make less-than-positive statements about VFP, and yet you felt they didn't really 'know' the product?

Quite likely you've run into that. Most of us have. Very likely you were upset, as I often was as well. So unless you want to fall into the trap that others have fallen into, my suggestion is to actually give the tool a decent evaluation before judging it.


Kevin