Sparse polyalgorithm seems to have a glitch

[ViralBShah]

As discussed here:

https://groups.google.com/forum/#!topic/julia-users/FagaqWVkBGU

Our sparse polyalgorithm is doing something wrong.

[ViralBShah]

Currently, factorize(S::SparseMatrixCSC) does the following (bottom of cholmod.jl), which is too simplistic.

# placing factorize here for now. Maybe add a new file
function factorize(A::SparseMatrixCSC)
    m, n = size(A)
    if m == n
        Ac = cholfact(A)
        Ac.c.minor == m && ishermitian(A) && return Ac
    end
    return lufact(A)
end

Instead, we should be doing as Tim Davis suggested in the thread on the mailing list:
http://faculty.cse.tamu.edu/davis/publications_files/Factorize_an_object_oriented_linear_system_solver_for_MATLAB.pdf

[ViralBShah]

#16 is the same thing.