An interesting debate. I'm currently one of the large number of aviators who fly something that cannot physically fit a power hungry, high RF Tx power system, so have a somewhat different perspective on this, perhaps. I've got lots of experience in bigger stuff and flew gliders maybe 30 years ago too, so can see this from most angles.
With the state of the art we have available at the moment, there is absolutely no technological solution that would even come vaguely close to giving a significant reduction in mid-air incidents. Far more than half of the private aircraft fleet in the UK are unable to carry and use a transponder all the time. Very few could fit a TCAS-like system. Unless transponder equipped aircraft become the majority, and they fly all the time with both transponders and radios on and serviceable, then we cannot hope to gain any real anti-collision benefit.
It's already been pointed out that the cost effectiveness of these expensive solutions doesn't stand up to scrutiny with respect to other ways of improving safety. £100M spent on better pilot training and skills enhancement would prevent many more incidents, I'm sure.
Mention was made earlier of the strobe detection system trialled years ago. Potentially this could be something that might be developed into a working system. We can now fairly easily detect strobes very reliably with cheap sensors, thanks to lots of work that's gone into missile detection systems over many years (and now largely unclassified).
Strobes are already commonplace, even on hang gliders and paragliders - I have a small one that fits to the top of my paramotor frame that cost me about £15. They also have the advantage of enhancing visibility as well as potentially providing a trigger source for an anti-collision system.
Although not a panacea, such a system would not require such a massive outlay as either a transponder based anti-collision system or something like Flarm. All that is needed is a cheap and reliable detector system, which would, potentially, be a great deal simpler to develop than any other option at the moment, in my view.
VP