The reason Smolin is not taken very seriously in the physics community these days is because he constantly expresses deep misunderstandings about key ideas, he calls research funding into question when he disagrees with how physics is being practiced, and he has contributed very little in the way of new ideas that have passed the stringent tests of the community. Unless you believe that there's an international conspiracy to shoot down his ideas, this fact alone implies that you must accept that the physics community does possess a healthy dose of scientific skepticism about proposals that simply don't work.
Indeed, Smolin seems completely detached from much of contemporary physics, from the phenomenon of decoherence to the modern understanding of quantum field theories in the framework of effective field theory. He seems to have no idea that there exists a perfectly acceptable low-energy effective quantum theory of gravity that has made bona fide quantum-gravity predictions (albeit astronomically small) for corrections to graviton scattering amplitudes and Newton's law of gravitation, indicating that the real problem lies not in fusing general relativity and quantum mechanics (indeed, general relativity is a prediction of the quantum mechanics of spin-2 massless particles), but in obtaining a theory valid up to arbitrarily high energy scales. The fact that Smolin has performed calculations in string theory means very little--one of the most amazing aspects of physics is how much you can calculate even if you have no idea what's going on. Just ask many of the physicists of the early 20th century, who thought for example that renormalization was just a "shell game" in which one "swept infinities under the rug."
How about Smolin's statement about how long it should take for a new idea to become a full-fledged, well-defined, and testable theory? Well, it took classical mechanics hundreds of years to go from Descartes to Newton. It took special relativity several decades from the time of Maxwell to the time of Einstein. It took GR only a few years after relativity--about ten. It took quantum mechanics over twenty years to go from Planck's first observed abberations in the black body radiation law and the ad hoc semiclassical approach to the modern formulation, but even then there were lots of mysteries that took subsequent decades to unravel. The Standard model was pretty quick--just four decades from the discovery of the neutron to the finishing touches on electroweak spontaneous symmetry breaking. String theory has taken about thirty years so far, but it's much larger and more complicated than anything that has come before it, so what's the big deal? Nobody can tell how long it takes a new model of nature to prove itself. And many string theorists have their hands in plenty of other approaches as well--they'd be quite happy to jump on a better framework if they or someone else discovered one.
How about Smolin's claim about whether a theory people are working on should make obvious predictions? Well, for quantum mechanics, it was literally decades before people knew what the real observables were, and whether you could actually test or predict anything, or whether determinism was lost and there was nothing you could use the theory to predict.
How about claims that a theory should be pursued only if it is forced upon us by observation? Indeed, some theories were pursued because experimental observations forced physicists to pursue them--nobody would have come up with quantum mechanics if experiments hadn't made it necessary. But general relativity was devised almost entirely by thought experiments in the total absence of experimental data--no, Einstein wasn't thinking about Mercury's precession at the time, nor did he have any data that the sun actually bent starlight by its gravity. General relativity was purely devised because Einstein though it was elegant and beautiful, and because it resolved some entirely theoretical inconsistences between special relativity and Newtonian gravitation, none of which had actually been seen in experiments up to then (except for Mercury). General relativity was as "pre-Baconian" as some people say string theory is, and seemed to address questions entirely removed from any practical use or human experience. Had it not been for complete and unlikely historical accidents like the existence of the planet Mercury or the fact that the moon and the sun happen to subtend the same angular size from the Earth's surface to allow for total eclipses, then there would not have been any conceivable experimental confirmation for the unimaginably tiny effects predicted by general relativity for many decades (today, we have devices like GPS satellites sensitive enough to detect the small discrepancies due to general relativity).
String theory isn't entirely to blame for what opponents claim is a failure to make testable predictions. The problems string theory is trying to address are problems of quantum gravity, and any theory that claims to address problems of quantum gravity is going to have trouble making testable predictions. Quantum-mechanical modifications to gravity are literally astronomically small. That's not a failure of string theory but a statement of fact about the kind of problems string theory is trying to solve. You can insist that these problems (like the origin and fate of the universe and the physics of black holes) are not worthy of attention, but you can't fault any model of physics that tries to address them for being difficult to test.
However, it should be noted that there actually is a huge source of data that constrain string theory; after all, string theory has to replicate the Standard Model. Indeed, if string theory's only accomplishment is to replicate the Standard Model at low energies (in at least one of its many vacua), explain the origin and nature of dark energy, and provide a consistent quantum theory of gravity at arbitrarily high energies, then it is certainly a better theory than the quantum field theory version of the Standard Model. It may not be the whole story, but it would certainly be an improvement. And most string theorists aren't claiming otherwise--as string theorists constantly say when speaking or writing for the public, they aren't betting their lives on string theory. String theory isn't religion--otherwise it would be a lot easier to raise money!