Ah, this is still on-going.
Suppose you were going to a point on the ground that was the same elevation as your departure point, it was not very far away, isobars nice and spread out that day in the immediate region, and there is no really close station to give you QNH at that point you are going to/by? - well what better could you use than the setting that gives you zero at your departure point? - the QFE?
Funnily enough, ZD576 had one subscale set to the QFE at Aldergrove and the HLS at waypoint A on the Mull had the same elevation :eek: Oh and the RADALT idea wouldn't have been much good with a sudden rise up the cliff from the sea to that HLS.

The radalt works perfectly in the context of my post, Walter. Such procedures are carried out hundreds of times every day in all kinds of weather around the world. The baro alt, with no exceptions that I know of, is set to QNH. I accept a radalt won't stop you flying into a cliff if you're intent on doing so.

Other than that I can't follow your logic. Sure, a QFE could be used as you describe. But so could a QNH without the requirement of being the same level. And this would also give your level relative to any spot ht (or cliff top!) on your chart.

<< The baro alt, with no exceptions that I know of, is set to QNH.>>
It is indeed a requirement that it is set to QNH when low flying en route, the exception being when you are imminently approaching a point at which you want to land or pass closely, then you may have the QFE set.
One of the settings was for the correct QNH but the other was appropriate for a QFE at the point in question.

I don't entirely follow that, Walter but in my sentence that you quoted I was referring to low-vis (CAT 2/3)approaches.

Why don't we all just use QFE for accuracy when looking to land. :D

ShotOne. Not sure the RAF does a lot of low-vis approaches. Ssuspect the post referred to poor wx/low ceiling/less than ideal vis rather than the formally defined cat 2/3.