I'm trying to figure out the P7 dependent parameters.

The manual, as so often, cuts short where it should not do so. The few odd pages of manual give no less than three different examples,  with most text not relevant and all incomplete. (sigh!)

I understand it is a listening method: one parameter can listen to another and be modified (by math addition) depending on the parameter connected to.
I understand that value additions for three or more values of the independent parameter, and that the actual additon is derived by parabolic interpolation. So far this is clear. The normal requirements to allow parabolic interpolation must be observed.

Question: Does anyone know how the reaction will be if the value of the independent parameter is outside the given table? Is the result then the nearest value in the input table, or is it a (parabolic) extrapolation and do I need to add extra points to prevent the result  jumping trough the roof? And then can I simply copy the last datapoint (assuming the interpolation routine is safe against division by zero), or do I need to observe a minimal (x) distance between datapoints there also?

As this is a new feature of P7, I'd suggest experimenting and reporting back (unless you want to wait for someone else to do it).

I reckoned that would be the way to go. I just wanted to avoid re-doing what somebody else did already, and maybe trigger someone at EF developers who knows.