1: \begin{abstract}

2:   This article introduces the DPG-star (from now on, denoted DPG*)

3:   finite element method. It is a method that is in some sense dual to

4:   the discontinuous Petrov--Galerkin (DPG) method.  The DPG

5:   methodology can be viewed as a means to solve an overdetermined

6:   discretization of a boundary value problem.  In the same vein, the

7:   DPG* methodology is a means to solve an underdetermined

8:   discretization.  These two viewpoints are developed by embedding the

9:   same operator equation into two different saddle-point problems. The

10:   analyses of the two problems have many common elements.  Comparison

11:   to other methods in the literature round out the newly garnered

12:   perspective.  Notably, DPG* and DPG methods can be seen as

13:   generalizations of $\mcL\mcL^\ast$ and least-squares methods,

14:   respectively.  \textit{A priori} error analysis and

15:   \textit{a posteriori} error control

16:   for the DPG* method are considered in detail.  Reports of several

17:   numerical experiments are provided which demonstrate the essential

18:   features of the new method.  A notable difference between the

19:   results from the DPG* and DPG analyses is that the convergence rates

20:   of the former are limited by the regularity of an extraneous

21:   Lagrange multiplier variable.

22: \end{abstract}

23: