Sir George, I'm not sure I understand your point. You seem, in places, to be blurring the two issues that I mentioned.
The first situation that you describe - airports with ATC and existing conventional ground-based navaids - satisfies both my a) and b). That these airports may not be pushing for satellite-based approaches is probably because they do not see a positive cost-benefit in doing so - i.e., it will cost money to get the approaches designed and the aircraft that they are providing with services can do very well with what's already there.
So that leaves us with airports without an approach control service or conventional navaids. The rub, as you put it, is that the airports will still have to pay to get the approaches designed but the users will be reluctant to pay for this through increased landing fees (or there simply isn't enough traffic to recover the cost in any reasonable period of time). I'm not sure that I see the lack of radar service or contiguous suitable airspace in which separation can be provided in the en-route phase to be in any way connected.
You say 'This means that a GA pilot wishing to land at an aerodrome with an RNAV(GNSS) IAP but not much else has to fly most of the procedure (whether IMC or not) in class G - i.e. not a known environment'. This is going to be true
in most cases - and is sometimes the case at airports with ATC and conventional navaid-based approaches. And, to me, seems to blur the issues of approaches and ATC/separation.
The mess I referred to is more to do with inconsistent and outdated rules which are commonly, if not ignored, creatively interpreted (and powers that be that have trouble recognising this). There's no question that satellite-based approaches offer great potential benefits but to meet the relevant ICAO SARPs may be difficult or costly for many airports. Maybe a better first step might be to try and develop some form of cloudbreak procedure which could be a more viable proposition in some situations.