Superstring theory has in foundations tremendous "leap of faith" - why it is dangerous and why other approximations may be work better.

 When superstring theory was started the scientists made not one but two important assumptions: second one is that the primal, most fundamental entity is one-dimensional (string) and first one, which is extremely general is that such a fundamental, smallest, non-dividable further entity exists at all. The first assumption is by far more important because there is  a lot of physicists (me included) who are sure that no fundamental entity exists at all, all the phenomena are emergent (made from something smaller), quark, for example will be split one day and any present day idea about the smallest entity is nothing more but the approximation, fit of nature laws, sooner or latter being replaced by something more fundamental, then by even more fundamental and so on.  

In this paradigm the theory of everything is impossible, scientists are discovering not the utmost laws but merely an approximations, fits of the natural phenomena doomed to be obsolete one day. Something like it was an approximate theory of liquid drop nucleus, which is now replaced by shell theory and will be replaced by something stemming directly from quarks one day and so on and so on. In this consideration the superstring idea is nothing more than  a fit of smaller but not smallest entity. That would be a bad fit for this situation. This is because once the even smaller than superstring entity is proposed, the interaction of those entities would try to minimize the energy, make short the surface and finally break the string. Null-dimensional primal entities (fit for right now) looks much better approach merely because they already have the surface energy minimized (sphere is the entity with smallest surface). If instead of superstring the superballs would be considered, they one day may generate at least some useful insights into the organization of the nature on the smaller level like liquid drop nuclei model was able to predict some useful phenomena like fission or oscillations of the nucleus [1-4].

This is because the analog of surface tension exists on all levels of the organization of matter - from round shape of planets, to the droplet of water, to the shape of nucleus - whenever even smaller entities are present, the inevitability of interactions between them demands decrease of total energy through round shape. If the interaction between entities are anisotropic the shape may be different (crystal) but in this case the existence of such anisotropy hints onto the lower level of matter organization present.

Only if the researcher is making a tremendous "leap of faith" and assumes that the entity is the smallest possible one may he or she consider it in a shape of string or bran (one-dimensional or two dimensional). In this case the shape may be different - by definition it is smallest, but if the assumption is wrong nothing useful will ever appear - the underlying entities will be assembled into something like balls (recall nucleus) not strings, because the string (cylinder) would have the surface energy well above minimum. The whole work will be a complete bullshit. That is why it is much safer to "play" with smallest entities being considered as null-dimensional - they will be eventually considered as built from something smaller but will work as a useful approximation [1-4]

References.

1.Semi-empirical mass formula - Wikipedia

2.Shape of the atomic nucleus - Wikipedia

3.The Discovery of Fission - Moments of Discovery

4.A comprehensive view of nuclear shapes, rotations and vibrations from fully quantum mechanical perspectives