Agreed. I never much cared for matrix algebra. But do you see the method?
1) Rewrite the inequalities as equalities using slacks (usually one slack per constraint equation). 2) Place in a rows/matrix form. 3) RREF the xyz (and the rest) as to get the xyz portion of the matrix as close as you can to a diagonalized matrix. 4) Write down the result algebraically from the (hopefully) diagonalized matrix and back substitute to isolate solutions. 5) Keep in mind, R=1 row is the maximized value.