The very interesting concept of relativity… so brilliantly introduced by Einstein and before him, by Galileo in some weaker form and by the ancient greeks in an even weaker form is an expression of the fact that we may chose the mathematical constructions we consider relevant and describe reality by their means. This is of course very general and it amounts at the property called “observability”… one assumed in the end that the things that are truly observable should not depend on the details we use to model them. Essentially there is some credit in drawing a picture of the trajectory of an electron, there is some merit in calculating its trajectory using electrodynamics… and both methods are equally valid if the results agree with the experiment. In some cases we may gain some extra insight in the problem we wish to describe, in other cases some insight is lost but another one is gained and so on. The same principle of “relativity” (in a lose interpretation) stays at the fundaments of quantum mechanics and probably string theory and who knows at what fundaments more… There are several descriptions one can go to, there are several concept one can introduce, they may have meaning in one part of reality and no meaning in the other but they are not reality. They are just means to describe it. Reality is whatever results and is observable. You may ask: but, is string theory not observable in principle? We just need a really big accelerator… well, that was the general belief in the 1980… namely that if we could look at the reality careful enough we would in the end see strings… This turned out to be nonsense. (well, there might be some who still believe this, but they are generally not “up to date”). What emerges now, slowly, is the theory of representations. This is an important concept, not very used in string theory but used in other fields of physics.