1: \begin{abstract}

2: A new method is introduced for solving Laplace problems on 2\kern .5pt D

3: regions with corners by approximation of boundary data by the

4: real part of a rational function with fixed poles exponentially

5: clustered near each corner.  Greatly extending a result of

6: D. J. Newman in 1964 in approximation theory, we first prove that

7: such approximations can achieve root-exponential convergence for a

8: wide range of problems, all the way up to the corner singularities.

9: We then develop a numerical method to compute approximations via

10: linear least-squares fitting on the boundary.  Typical problems are

11: solved in ${<}\kern 1pt 1$s on a desktop to 8-digit accuracy, with

12: the accuracy guaranteed in the interior by the maximum principle.  The computed

13: solution is represented globally by a single formula, which can

14: be evaluated in a few microseconds at each point.

15: 

16: \end{abstract}