Christof....

1 quick point... it looks like you have some HTML tags in the text of your message.


>Set Assert off
>
>nStart = Seconds()
>For t=1 to 100000
> Assert .T.
>Endfor
>? Seconds() - nStart
>
>nStart = Seconds()
>For t=1 to 100000
>Endfor
>? Seconds() - nStart


I don't think this is a good test. Continually invoking assertion tests in a loop is not something that would normally be done IMO. Rather, I would see something like this as being more in the norm:

lParameters x,y,z
Assert Vartype(x) = "C"
Assert Vartype(y) = "D"
Assert Vartype(z) = "N"

x = upper(x)
y = dtoc(y)
z = str(z)

return


>The loop with Assert takes between 40% and 50% longer than the loop without any command in there. The first loop need averagely 50ms, the second 35ms on my system. This is independant on whether this code is executed in the development environment or with the runtime library.
>
>>Christof, I can't emphasize enough how good it is to see you up here.
>
>Well, it's still just a visit. Although someone in the UT doesn't seem to beleive me, the envelop beside my name is now blue and I'Ve access to functions I hadn't before. :)

What the heck does that mean? Are you going to be saying that in another year??? :) I think the overall content up on CIS is still better. The dialog you and I are having here is typical of CIS and what makes CIS good. UT however is quickly catching up.