The 16 members of a 4x4 matrix are usually numbered in mathematics like Y and X coordinate (the university teachers speak of columns and rows, which is kinda stupid, since even I think in Excel with x and y coordinates (much easier to make VB scripts that way)).
So when they say m11,m12,m13,m14 (or m00,m01,m02,m03, if they start with 0 instead of 1, which is unusual since when talking about columns and rows you start with 1, so basically mathematicians have no clue about consistency, only C++ programmers do), they actually mean: cell at coordinate 1,1; 2,1; 3,1; 4,1. So in LE you have TVec16 which calls them A0,A1,A2,A3. For y==2, it's called B0,B1,B2,B3, etc...
I could have called them like in mathematics, but then I would have needed to use y,x=x,y coordinate system, which would cause even more confusion between programmers and mathematicians. So calling them with letter+index can not be interpreted wrong, and it's also shorter to type. And yes, it is a common standard notation also, I checked several sites about it. It also unifies vector and matrix notation which is kinda cool.
And since it's a Vec16, and not a Mat4 (which are the same thing though), you don't use x and y coordinates in Vecs. If I would have used Mat4, it would have been inconsistent again since it would be totally different that the Vec2,3,4,6 structs, which I wanted to avoid, since it's basically the same thing: just a container for an n-amount of floats.
However, in addition to TVec16, there could be also TMat4 in the C headers, which would be the same as TVec16, but only using different member names. How useful that would be, I don't know.