Name of h is misleading

[ramcdougal]

@Helveg observed that

>>> from neuron import h
>>> h
<TopLevelHocInterpreter>

While h can be used to interpret Hoc, that is not its primary function. It's an interface to the NEURON stack machine at a level below the Hoc interpreter. (I tell people that h stands for NEURON, just like d stands for integer in C string formatting.)

Given that NEURON 9.0 is coming up, maybe now is a good time to change how it displays. What about:

>>> h
<NEURON>

or

>>> h
<NEURONStackMachine>

I'm not picky, but there's no reason for anyone to see "Hoc" here.

Unfortunately, I think we are stuck with

>>> type(h)
<class 'hoc.HocObject'>

because people may be subclassing from that, but I don't see anyone making assumptions that depend on repr(h)

[nrnhines]

I don't have an objection (or at least any objection would be very weak) to

>>> h
<NEURON>

In favor, h provides access to "all" of NEURON. The weak objection I suppose is the relation of that to the already existing
top level module name import neuron and the subsidiary neuron.nrn vs neuron.hoc modules. I've always tried to distinguish between general interpreter features and neuroscience specific features. That dates back to the code folder names, oc and nrnoc. (Long ago, when object notation was added to hoc, the hoc folder was renamed to oc). At the moment, in neuron/__init__.py one can see that

h = hoc.HocObject()