Dear Mozibur Rahman Ullah
I really don't think special relativity is simple. The Lorentz transformation is simple; the theory, based on the 4D ontology is complex and paradox-ridden.
You say, "special relativity shows its subtlety when we try to incorporate it into other theories, that is, into gravity and into quantum theory."
But in a recent paper Glavan and Lin note that, "According to Lovelock's theorem, Einstein's general relativity with cosmological constant is the unique theory of gravity if we assume (i) the space-time is (3+1) dimensional [plus three other conditions]."
The field equations are not Lorentz invariant. If one tries to add the Lorentz invariant pair-wise connections between all local particles in a global gravitational ontology, one would be adding pair-wise distortions of time and space (length contraction and time dilation), all local-pairwise-velocity-dependent, onto global mass-dependent space curvature. For example the Schwarzschild metric is time independent; it is frozen in space forever. The whole thing is an untenable proposition.
Newtonian gravity is Galilean invariant. Recall that Einstein's field equations must make contact with the Newtonian potential in order to be a physical theory of gravity.
The Maxwell-Hertz equations are Galilean invariant -- Einstein based his theory on Maxwell-Hertz, but he mistakenly used (ch 13), "bodies at rest" and he Lorentz-transformed between two cartoon worlds, whereas (ch 14), "bodies in motion", with the convective derivative is Galilean invariant.
Schrödinger quantum mechanics is Galilean invariant, not Lorentz invariant.
When Dirac forced special relativity symmetry on his equation, he gets a free particle with speed 1.7c, that is faster than the speed of light. When his equation is used for particles interacting with the field, he loses the 'spacetime symmetry' he had forced.
Feynman, as best I can tell, uses Lorentz to maintain inertial mass, and ignores any length contraction issues, since these aren't measurable. The inertial mass can be maintained without Lorentz. However I suspect that the enforced geometry group symmetry probably simplifies a lot of the math in path integration, etc.
In short, only special relativity, a toy model without gravity or rotation, or even inertial mass, requires Lorentz, and that is because Einstein invented a universal time dimension for every cartoon world, and 'attached' a constant speed of light to every world to enable the Lorentz to be derived, and used the inappropriate Maxwell-Hertz equations.
I do not think Lorentz has travelled at all well across theories, but these facts are not always pointed out.
I do thank you for reading my essay and thinking about it.
Warm regards,
Edwin Eugene Klingman