I am trying to prove that $f(x,y,z) = x^2 + y^2 + z^2 - 2xy - 2xz - 2yz$ is irreducible over $\mathbb{Z}[x, y, z]$ (I am sure it is) and maybe over $\mathbb{R}[x, y, z]$ (almost sure).
I checked that it is not divisible by $x+y+z$ so factors are not symmetric. And that's where I get confused. Suppose $f = g \cdot h$, then some permutations must swap $g$ and $h$ and maybe there are some that preserve them. How do I make sure I checked all the possibilities? Is it sufficient to check what happens if $x,y \mapsto y,x$ swaps them and $(123) \in S_3$ swaps them?
If the form factors: let it be $$ (ax+by+cz) \cdot (dx+ey+fz) $$ in which $a,b,c,d,e,f$ are permitted to be real or complex.
Multiply it out, take its second partial derivatives and arrange in the Hessian matrix,
$$ H= \left( \begin{array}{ccc} 2ad & ae+bd & af+cd \\ ae+bd & 2be & bf+ce \\ af+cd & bf+ce & 2cf \\ \end{array} \right) $$
Finally, calculate the determinant, which is $$ \det H = 0 $$
So, if the form factors, the determinant of the Hessian is zero. Contrapositive, if that determinant is nonzero, the form does not factor.
Your Hessian matrix, rather half of it, is $$ \left( \begin{array}{rrr} 1 & -1 & -1 \\ -1 & 1 & -1 \\ -1 & -1 & 1 \\ \end{array} \right) $$ and its determinant is nonzero