Inconsistent output from time response routines about python-control

[murrayrm]

If memory serves, this is on purpose. The idea at the time was that we wanted the output of step_response (and similar functions) to work better with plotting. In the current pattern, plot(step_response(sys)) works (I think).

The idea with the matlab module is to map everything to the "conventional" MATLAB terminology. But for python-control itself, we can do things differently if we want.

If we change this, we need to bump the version number from 0.6.x to 0.7.x, I think. We will break any existing scripts that people are using (unless they just import the matlab module).

We should also compare to scipy.signal, which I think has some response functions.

[cwrowley]

I agree that plotting syntax is nice. I also just checked the scipy.signal routines, and they also use the syntax t,y (opposite of Matlab), so I agree it's better to stay consistent with scipy than Matlab.

I'll leave this issue open, though, since we should still fix the docstrings for the routines from matlab.py.