1) new experimental data which conflicts with the currently accepted laws of physics
2) inconsistencies within the current laws of physics
or
3) The "error" in physics is not there and there is an error in the derivation of the inconsistency.
What happens here is that an engineer attempts to use the accepted laws of physics without measuring new experimental data in hopes of violating the accepted laws of physics, but without formulating an explicit contradiction. This is why gus_massa's BS alarm got triggered: using only accepted laws in physics it should be impossible to derive a contradiction, and any seeming contradiction is either a complicated form of already understood propulsion mechanisms (photon rockets) which consume rocket rest mass, or else the seeming contradiction is an actual contradiction, in which case the derivation should be simplified to the extreme to point out the inconsistency in physics. The latter was not illustrated so it entails either an error in derivation, or an inefficient photon rocket.