doc: Review and update documentation for "Matrix Manipulation" chapter. * matrix.txi, arrayfun.m, blkdiag.m, fliplr.m, flipud.m, logspace.m, postpad.m, prepad.m, randi.m, repmat.m, rot90.m, rotdim.m, shiftdim.m, sortrows.m, vech.m, xor.m, hadamard.m, hankel.m, hilb.m, invhilb.m, magic.m, pascal.m, rosser.m, sylvester_matrix.m, toeplitz.m, vander.m, wilkinson.m, bsxfun.cc, find.cc, lookup.cc, rand.cc, tril.cc, data.cc, arrayfun.m, blkdiag.m, fliplr.m, flipud.m, logspace.m, postpad.m, prepad.m, randi.m, repmat.m, rot90.m, rotdim.m, shiftdim.m, sortrows.m, vech.m, xor.m, hadamard.m, hankel.m, hilb.m, invhilb.m, magic.m, pascal.m, rosser.m, sylvester_matrix.m, toeplitz.m, vander.m, wilkinson.m, bsxfun.cc (bsxfun), find.cc (find), lookup.cc (lookup), rand.cc (rand, randn, rande, randg, randp), tril.cc (triu), data.cc (all, any, horzcat, vertcat, cat, permute, ipermute, ones, zeros, eye, linspace, resize, reshape, issorted, diff): Improve docstrings

--- a/scripts/special-matrix/wilkinson.m
+++ b/scripts/special-matrix/wilkinson.m
@@ -20,10 +20,10 @@
 ## @deftypefn {Function File} {} wilkinson (@var{n})
 ## Return the Wilkinson matrix of order @var{n}.  Wilkinson matrices are
 ## symmetric and tridiagonal with pairs of nearly, but not exactly, equal
-## eigenvalues.
+## eigenvalues.  They are useful in testing the behavior and performance
+## of eigenvalue solvers.
 ##
-## @seealso{hankel, vander, sylvester_matrix, hilb, invhilb, toeplitz,
-##          hadamard, rosser, compan, pascal}
+## @seealso{rosser, eig}
 ## @end deftypefn
 
 ## Author: Peter Ekberg