Inconsistent behaviour when differentiating MatrixSymbols with respect to each other

[r0uv3n]

Assume we have

import sympy

n = sympy.symbols("n",positive=True,integer=True)
X=sympy.MatrixSymbol("X",n,1)
Y=sympy.MatrixSymbol("Y",n,1)

Then sympy.diff(X+Y,Y) return I (which is consistent with e.g. the behaviour for normal symbols, where sympy.diff(x+y,y) returns 1), but sympy.diff(X,Y) returns Derivative(X,Y) (for normal symbols, sympy.diff(x,y) returns 0).

It seems sympy assumes a dependency between X and Y if we the expression does not explicitly depend on Y, but not otherwise.

Is this a bug or am I misunderstanding how this should work?

Right now I really just want to get the coefficients of an expression of the form a*X+b*Y (which will always be linear in X and Y), for which I'd like to use .diff(Y) (since .expand().coeff(Y) also doesn't seem to work). Is there some other way to achieve this without running into the issue above?