Dear developers,

I am a happy user of Matplotlib,  which is a wonderful library - thank you!

Would it possible to support the SciPy Array interface in more places to make
Matplotlib less dependent on Numpy/numarray?

In Ubuntu Edgy's version of Matplotlib it's possible to use CVXOPT matrices
in some places, but not in others:
>>> from cvxopt.base import matrix
>>> import pylab, numpy
>>> pylab.imshow(matrix([[1.0,2.0],[3.0,4.0]]))
<matplotlib.image.AxesImage instance at 0xb5d560ac>
>>> pylab.matshow(matrix([[1.0,2.0],[3.0,4.0]]))
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "/usr/lib/python2.4/site-packages/matplotlib/", line 1431, in matshow
    w,h = figaspect(arr)
  File "/usr/lib/python2.4/site-packages/matplotlib/", line 932, in figaspect
    arr_ratio = float(arg)
TypeError: float() argument must be a string or a number