c2d of improper transfer function

[mfalt]

c2d of improper transfer function breaks silently since it goes through state-space, and isproper is broken, see #85
Should we enable some way of discretizing a transferfunction other than going through state-space form?

[olof3]

Yes, at least it should be possible to do pole-zero matching for systems on zpk form.

[mfalt]

Isproper no longer breaks silently, but this feature should still be added.