In numerical modeling, a numerical computational scheme is said to be consistent if the discrete algebraic equations created by the process of discretization recover or reduce to the original continuum differential equations as the spacing in the computational grid is shrunk to zero. The scheme is said to be unconditionally consistent if the above is true no matter how (i.e. in what order, etc.) the grid is shrunk. Thus consistency deals with relations between equations in their continuum versus discrete forms, as opposed to convergence.