>>Merry < whatever >! :)
>>
>
>Right back at you <g>
>
>>Did you include the time to instantiate the object?
>
>No, I excluded that from the test.
>
>>Build an exe with a class and a udf.PRG. You don't need to SET PROCEDURE TO or SET CLASSLIB TO to execute the udf.
>
>Very interesting! I wonder what VFP is doing differently for an EXE? Now it's around 45% in favor of the UDF. The time to execute the class call hasn't changed much in the EXE (it dropped slightly), but the function call has dropped significantly.
>
>Cool.

No offense, but this is one of the things I'm surprised more people don't know. If you build stuff into an exe there's better performance than if they're left on the disk. It's not just the difference between the OS searching for files. It seems VFP has a special way to access files in the exe. I'm not sure of the mechanism, but the result have been there for at least 10 years! :) I distinctly remember showing this to Jim Nelson when we first worked together.

Performance testing of the final exe is best!