Sure, but I've always viewed Ramanujan's summation of divergent series to be a curio, I'd never have expected to see it turn up in physics. Pondering this a little, I think what disturbs me about this is the irreducible discontinuity it introduces. If you were to tell me that because of reasons, F = ma is wrong, it should have been F = m^{1.00000000001}a all along, then the resulting model of physics doesn't change that much; but if the Casimir equation should use ζ(-1+ε) rather than ζ(-1), then the sum is finite but as large as you like and the smaller that ε is, the larger this term is.