vertexEdgeMatrix and vertexFacetMatrix seem not to be needed. Also they print indices in the output matrices?

documentation:
*  statePolytope is documented wrong. It produces more output.

rays has weird definition
saveSession in own file?
polar is now defined via dual of underlyingCone. Will this fail for having linealitySpace?

generate polymake tests for:
*  hilbertBasis (DONE)
*  minkowskiSum (this used to ignore the lineality space completely?) (DONE)
*  normalFan (DONE)
*  faceFan

Thoughts:
*  Sanitize input matrices at some point?
*  There are too many intersection methods.
*  Completeness of fans is probably very expensive to check at the moment.
*  What is a better naming scheme for the properties? Should we adapt the
   polymake scheme?

Packages that use Polyhedra:
[lars@sunfish packages]$ grep -R "Polyhedra" . | grep needsPackage
./MonomialAlgebras.m2:needsPackage "Polyhedra"
./Polyhedra2.m2:needsPackage "PolyhedralObjects"
./SRdeformations.m2:needsPackage "Polyhedra"
./LatticePolytopes.m2:needsPackage "Polyhedra"
./supplanted-packages/Polyhedra.m2:     F << "needsPackage \"Polyhedra\"" << endl;
./Macaulay2Doc/test/schenck-book-2.m2:needsPackage "Polyhedra"

Packages that should go away:
*  PolyhedralObjects.m2
*  Polyhedra2.m2
*  polymake.m2 or polymakeInterface.m2

Undocumented methods, that have been removed:
*  newMinkSum: No reference for algorithm given and too long to fix.
*  dualCayley, dualCayleyFace: Not documented at all, never tested.

Methods without tests
*  stellarSubdivision
*  cellDecomposition
*  faceFan
*  polarFace
*  Everything concerning PolyhedralComplex
