Everybody knows that space-time has to go, but the reason discrete space-time short-distance grains are not discussed is because the naive idea are all wrong. If you give a heuristic space-time foam, you can get hand-waving justifications for finite theory, but there is no real argument there, because the path-integral for quantum gravity is a mess.
You can try to define the path integral as a sum over different manifolds, but how do you do a sum over manifolds? There are problems of principle, like the fact that reducing the sum over topologies counting each topology only once is known to be computationally undecidable. So you need to cut off the integral somehow at short distances so you can't have arbitrary handles pop up. Then perhaps you can make sense of the integral, but how can you cut off small distances when the distances are dynamical?
The path-integral is a dead end without new ideas. One such idea is loops, and this gives a direct picture of space-time graininess by spin-foams and things like that. But this idea is very hard to complete.
The best way out of the impasse is to change perspective completely. You look at the experimental questions you can ask. If you want to know what is happening in a microscopic region of space, all you can do is shoot particles at it, and get some particles out, in a quantum mechanical way. So you ask what S-matrix is happening there? That's the complete description of space-time.
This question sidesteps the issues completely, because the S-matrix is defined asymptotically on the boundary of the original space, using plane waves at infinity, so it is well defined using only the asymptotic symmetries of the big space. It doesn't make any difference if there is microscopic foam, or a big mess, or whatever, the S-matrix is well defined and stable to quantum gravity. This is why it is a good foundation.
So you look for an S-matrix theory, and this is string theory. Within string theory, you can find formulations in which you can see the space between black holes built up from the oscillations of those black holes. This is very close to a full atomic description of space. The quantities you have are only quantum mechanics in 0 space and 1 time dimension, and they build up a full 11 dimensional reality in the limit of many atoms.
The atomic description in string theory has no simple space-time description, because it is asymptotic, it respects the holographic principle. You can't say that space has been chopped up into volumes, because it is really only the boundary of space that needs chopping up. The interior is just a reconstruction. These ideas supplant more primitive notions of space-time graininess, and really, those old ideas are obsolete.