Proponents may, and do, claim pretty much anything. Until one of them actually produces a working machine, their claims are utterly irrelevant.
n the first instance, all
anyone can have is a "claim".
Belief in the claim may be based on testing performed in a lab with visiting world authorities, peer review papers and Nobel prizes all round. Or it may be based on a personal experience with God and a litre of Mezcal. It's still just a "claim" as far as the next test goes.
Partially working healing? So you can only heal a broken leg halfway? Better than nothing. Partially working perpetual motion? That's what is normally just called "motion". It's really not anything special.
Yes, all right mister smarty-pants - I am being imprecise. You may shoot me now.
I had hoped that my example (need frictionless bearings to make the machine go, cannot afford bearings, therefore do not have working machine) made my meaning clear.
If I say "I missed the target" and "I almost hit the target"... these statements mean the same thing. However, there is a different sense about them. The second indicates that the shot was in some way a "near" miss.
Similarly, partial pmm, while not actually pmm, is also not a brick (also not pmm).
If the mechanical too-much-friction wheel in the example was verified to have an efficiency of 1.001, once mechanical friction is accounted for, then I have demonstrated pmm in principle. What remains is to engineer sufficiently low friction bearings (ceramic, magnetic, whatever).
Of course, I'd have to demonstrate that it is indeed mechanical friction which is slowing the thing down - like: providing full access to the machine, providing detailed math, that kind of thing - and not some fundamental limitation in nature which boils down to a known conservation law.
I don't know, perhaps
you'd call that a working pmm? But it is not motion in perpetuity.
My thesis here is that: there exist circumstances in which it is possible to demonstrate the validity of a devices operation without having a working device handy.
In the particular case of pmm:
http://www.lhup.edu/~dsimanek/museum/test-pm.htm