I have a question regarding the calculation of right and left ideals of a matrix ring. I understand the concept, or so I believe, but, when consulting the argument on this page http://mathworld.wolfram.com/RightIdeal.html I harbour some doubts I would like to share with you. The calculation of the matrix product proving the right ideal is crystal-clear to me. However, when dealing with the argument for excluding the possibility of that very same structure being a left ideal, there is something I do not fully understand.
1) Does it mean the matrix belonging to the Ideal, when being multiplied by another matrix, has to occupy the inverse position to the kind of ideal we are dealing with (so that if we are talking about a right ideal, it will be the left-sided member of the noncommutative matrix product, and when talking about an alleged left ideal it hast to be on the right of the product)?
2) If 1) is the way I suggest, and given the counterexample in (6) on the link, is it the case, as I suspect, that by inverting the order of the product members in (6) we once again get an ideal, namely a matrix with all entries zero?
Could you please confirm me these points?
Thanks in advance.