Coverage remained the same at 93.591% when pulling f06ef73 on roryyorke:rory/iosys-repr into e8db508 on python-control:master.

I agree that a change would be useful here. More generally, it would be nice to think through what we want to have displayed in various situations:

• __repr__: this is the default display for objects in iPython/Jupyter and is supposed to be something that could, in principle, be used to recreate the object. For StateSpace systems, for example, this currently gives

StateSpace(array([[ 0.,  1.],
       [-1., -2.]]), array([[0.],
       [1.]]), array([[1., 0.]]), array([[0.]]))

• __str__: this is what you get when you type print(sys) and it gives a more verbose output. The current output for a state space systems is:

A = [[ 0.  1.]
     [-1. -2.]]

B = [[0.]
     [1.]]

C = [[1. 0.]]

D = [[0.]]

• After PR I/O system enhancements #710 gets merged, creating a system with ss will produce a LinearIOSystem, which has a different __str___ representation:

Object: sys[0]
Inputs (1): u[0], 
Outputs (1): y[0], 
States (2): x[0], x[1], 

A = [[ 0.  1.]
     [-1. -2.]]

B = [[0.]
     [1.]]

C = [[1. 0.]]

D = [[0.]]

• We also have a _repr_latex_ method, which gets used in Jupyter notebooks. This produces the following output:

Finally, PR #710 has a conflicting change since I reset the repr format there to include the type and system name. So an InputOutputSystem object has output:

<class 'control.iosys.LinearIOSystem'>: sys[0]

I like the version in this draft PR better, but note that if you get something with lots of inputs and outputs, it could be a bit odd looking.

PR #721 has a portion of this (__repr__) implemented, but leaves the other issues touched on above unchanged. It would be nice to sort out a consistent set of representations that make sense in different environments (eg, Jupyter vs command line).

FYI, Spyder (a nifty little IDE) also uses the _repr_latex_() output
It sure is handy to get your SS in a compact matrix in such editors.

This has been implemented since v0.10.0.