I think that's a bit too harsh an assessment. If a self-consistent theory can be devised that elegantly describes nature, it can be used to suggest ways to verify or test it. Of course, if the string doctors insist on proposing variations which are in principle untestable, then no one should take them too seriously.

The astrophysicists have similar problems with dark matter, dark energy, and all of the weird cosmological conjectures which have breen proposed lately; but as long as they can lead to observations which can be used for verification, the theories are useful even if dead wrong.

It is my understanding that some variants of string theory have already had to be constrained because they failed some experimental tests. That's progress, of a sort. And the experimentalists should be working on new techniques too. Theoreticians can be flaky, at times, but we need them; observers are great at keeping you honest, but they're not always the most creative ones out there.