Giotis has pointed out an argument published in the Guardian:
The title of the article asserting that "physics that has gone too far" is a rather accurate picture of the basic nature of the string theory's critics – their proximity to the Inquisition that wants to dictate which boundaries science isn't allowed to surpass. Science is "allowed" to surpass any boundaries. Any question where a sufficient body of evidence and relationships between the known and hypothesized facts may be developed is likely to become a fruitful subdiscipline of science. String theory undoubtedly belongs to this list.
Baggott effectively claims that string theory was completed 40 years ago and by this time, everything should have been settled. Duff explains to him that string theory was actually born 40 (well, 45 if one includes the years in which it wasn't called string theory but it was already the same subject) years ago and the insights have been accumulating during the subsequent years and they're still being accumulated which is really why the ongoing research makes sense.
Duff can't resist to point out an important example of such a development that occurred much less than 40 years ago – the discovery of 11D M-theory and the role of membranes in M-theory. It was important, indeed, and Duff himself has made important contributions to this advance years before M-theory got its name. My estimate is that a dozen of advances are as important as the discovery of M-theory and its connections to string theory.
Baggott, like many laymen, believes that science is supposed to talk about things that have already been measured only. Duff corrects him and explains that an important, nearly defining part of science is about predictions and analyses of things that have not been observed, even things that are not gonna be observed soon according to any plan.
So there's some debate in the British daily on whether or not the atomic theory or Einstein's papers about entanglement were physics or metaphysics or speculation. It's a matter of terminology but it seems totally obvious to me that this activity is an important part of what physicists do, have always been doing, and have to do for the progress in physics to be balanced. Einstein was clearly doing physics when he was writing EPR-style papers. We cite him (and them) for those insights and questions. Some of the physics was wrong, some of it may be classified as speculations about alternative theories that can't exist, but in some of it he just followed the proper laws of quantum mechanics to derive certain interesting phenomena (whose validity and whose character in Nature Einstein often misinterpreted and mispredicted).
It was right or wrong but it was surely physics and we cite Einstein for these things, too. What string theorists are doing is much more robust and rooted in the empirical data than Einstein's work about entanglement. String theorists are arguably much more right about string theory and questions it addresses than Albert Einstein when he talked about entanglement. Duff also has to point out that string theory is as rooted in the known empirical facts as the Standard Model.
There are other topics – I endorse every word by Duff – and at the end, Duff tries to clarify a common laymen's (and Baggott's) fallacy. They love to confuse implications derived from a theory and theory's assumptions. In particular, Duff explains that supersymmetry, extra dimensions, the existence of string or membranes (at energies smaller than or equal to the Planck scale), and various other types of physics aren't assumptions of the theory we have but its predictions, something we have derived from a more coherent and fundamental starting point.
Even if you're confused about these matters, you should be able to understand that it's important to distinguish assumptions from derived predictions – the difference is as important as the difference between positive and negative numbers, between credit and debit. They're really standing on the opposite sides of a seesaw so if you're confused which side, it is likely to dramatically invalidate your conclusions. ;-)
String theory allows us to derive known aspects of physics such as general relativity, gauge theories, chiral fermionic matter, but also new aspects such as supersymmetry, grand unification, exotic heavy states – and more theoretical insights such as the nontrivial mechanisms that preserve the information when black holes evaporate, and so on, and so on – from a more unified and fundamental starting point. Only the starting point which is extremely robust or rigid is "adjusted"; all the other claims are implications of them.
Update: There was another exchange in the Guardian on Sunday between John Butterworth and Mike Duff (e.g. first comment), among others. Butterworth isn't a layman. He is an experimenter but it seems he completely misunderstands the work of theoretical physicists. Among experimenters, he isn't the only one. I won't spend more time with Butterworth's musings because they're just a softly formulated form of the very same delusions that the deeply misguided anti-theoretical-physics mob writes everywhere. I sort of believe in the renaissance man so even if someone like Butterworth is an OK experimenter, his Fachidiocy and a complete misunderstanding of a closely adjacent discipline to his – theoretical physics – is just stunning. If he could at least keep his mouth shut instead of boasting his delusions.