l2 error norm of solution

Hi, I didn't find any built-in approach to estimate the error norm of solution (e.g., l2 norm) on analytic test. So I wrote my own functions for estimation of this norms, using barycentric integration on triangles and trapeziod integration on boundary edges. The question is if I understand correctly the connection between solution u and [p,e,t] mesh format or not. I used hyperbolic solver for scalar case, so I have solution u as in manual:

        For the scalar case, each row in the solution matrix U1 is the
        solution at the coordinates given by the corresponding column in
        P. Each column in U1 is the solution at the time given by the
        corresponding item in TLIST.

Am I right, that to assign some values of u on one edge only, I should use the following syntax:

      for i = 1:ne
          if e(5,i) == 4
              for j = 1:n
                  u(e(1,i),j) = ...;
                  u(e(2,i),j) = ...;
              end
          end
      end

Is e(1,i) refers to p value in this point?

[Dmitry]

Or may be someone already wrote matlab functions to estimate the error norm of solution? Please share here or give link to.