yes, well, and that's exactly the point I've been trying to make. Science is useful because it accepts things that are useful to accept, not because it accepts things as unassailably true, which is the commonest ordinary meaning of "proof".
Well, no, but I think it's worthwhile to first stop for a moment and consider what we mean by "useful" in this context. Scientific principles are useful to us because they illustrate or illuminate some aspect of actual reality. Whatever else it may be, science is not a solipsistic exercise, and is predicated on the assumption that objective reality exists beyond ourselves. Given that, for our principles to be truly useful in a scientific sense, they must then bear some relationship to objective reality, and hence be in some sense, truthful, albeit as an imperfect approximation of truth, rather than being provably and indisputably true. The inverse-square law of gravitation is not deductively and provably true, but it is useful to us precisely because it appears to bear some relationship to the actual, objective behavior of physical objects. Insofar as that behavior objectively exists and is objectively real, the inverse-square law must then be at least an approximation of that objective reality, else we would not find it useful. And it seems to me that the better those approximations are, the closer they are to that unnattainable objective truth, the more useful they then are in a scientific sense - i.e., there are degrees of usefulness which correspond in some sense to the principle's correspondence with external reality.
This should not, of course, be taken to mean that "useful" and "true" are equivalent, or that "useful" even necessarily implies "true" in all cases. But I think it's clear that the principles we find useful are useful precisely because they represent some approximation, whether near or far, of actual reality, and hence, truth. When someone is sent up for trial, we expect the state to prove its case, that the defendant is to be proven guilty. Clearly, however, we do not expect the state to conclusively prove its case true as a matter of logical certainty. Instead we expect proof beyond a reasonable doubt, proof as a matter of reasonable certainty rather than absolute certainty - after all, no matter how good the prosecutor is, he or she can never eliminate as a matter of certainty the hypothesis that Nicole Simpson was killed by Colombian drug dealers or Martians or Santa Claus or whatever.
Instead, it is "proof" in an inductive sense that we seek, not proof in the absolute deductive sense. And I think this is the sort of proof that science is generally concerned with, where we examine where the bulk of the evidence rests, and when the evidence overwhelmingly lies in one particular direction, we then reasonably conclude that the issue is, at least for the moment, proven as a matter of reasonable certainty. Which we can do precisely because we don't restrict "proof" to mean "unassailably true" most of the time. In fact, by "proof", we usually mean "reasonably certain", and it's hard to see how it can be any other way - if we reserve the word "proof" for only those situations deductively demonstrable, we will probably tend to find that the word "proof" has ceased to have much utility for us at all, having no real application to any circumstances beyond the purely hypothetical.
It is certainly fair to state that the theory of evolution is not conclusively proven true, but I do not think it is at all unreasonable to state that, at least for the time being, it is proven as a matter of reasonable certainty. The theory of evolution is not expected to be taken axiomatically, after all - we consider it to be true because that is the direction in which the bulk of the evidence currently points. In the sense that it then illustrates some aspect of objective reality for us, it is both useful and reasonably true.