An effective way of computing bounds for inverses of linear operators is discussed. It involves computing very fast a bound for the inverse of linear Volterra and Fredholm integral operators. This approach covers a wide class of ordinary and partial initial and boundary value problems. It is demonstrated that this concept is feasible in the framework of computer algebra systems.