As I recall, Rovelli never published anything about DSR until August 2008, when he posted "A note on DSR" on arxiv (0808.3505). And he never took part in the discussion of whether GLAST spacecraft (now called Fermi) might or might not observe that high-energy photons are delayed. It appears that he had a speculative note suggesting a possible physical mechanism which might cause DSR to appear (as an approximation to reality) in the limit. And this note circulated privately for some time before he posted it in 2008. I can appreciate the caution in dealing with a speculative risky idea that one isn't sure about. However this risky idea is interesting, I think. You might like to consider it. Keep in mind that from the standpoint of a relativist special rel is not fundamental. SR only arises as an approximation in a certain flat limit where G -> 0, or where matter is negligible. What is fundamental is the dynamic geometry of GR, which is the cause of any flat approximate geometry that one might find applicable in particular situations. Or more precisely it is the sought-after quantum version of geometry that is fundamental--the quantum GR. So one can imagine taking limits from that theory, as hbar -> 0 or/and as G -> 0. What one gets depends on how one takes the limit, and will determine for which application the approximation is appropriate. Rovelli asks, what if we take limits as hbar and G both go to zero but the ratio hbar/G stays constant? Well hbar/G is the square of the Planck mass. So that amounts to keeping the Planck mass constant while one lets G -> 0 (the flat empty space limit) and hbar -> 0 (the non-quantum classical largescale limit). Curiously, it looks like he gets something like DSR when he does this. The point is that there is some physics which does not depend on hbar separately or on G separately, but which depends on the ratio hbar/G. And this carries over into the limit. So if you take the limit in a clumsy way where you first let hbar -> 0 and then you let G -> 0, then you miss seeing these effects. You have to take the two limits simultaneously, keeping the ratio constant, and then you see the effects. The paper is only 6 pages, and comparatively simple writing, so you can see for yourself http://arxiv.org/abs/0808.3505 how it works out. Better than if I try to repeat every step of the argument. Just a comment: there is a curious epistemological point that comes out which is that since points in space or spacetime have no objective reality you cannot define a reference frame without designating some particle or particles of matter. So it is philosophically necessary that a reference frame must have a mass associated with it. This is quite strange. At least to me it never occurred to think this before.