Use defaults to find parameter_names

[Christian-B]

As we are moving to using spynnaker/pyNN/models/defaults.py for all models

Why not have that class also discover the data needed for
AbstractHasParameterNames.get_parameter_names

This will avoid issue like in

sPyNNaker/spynnaker/pyNN/models/neuron/plasticity/stdp/timing_dependence/timing_dependence_vogels_2011.py

Line 46 in 64d0914

__PARAM_NAMES = ('alpha', 'tau')

Where some params are included and others are not

[Christian-B]

PyNN specification divide the init values init default_parameters and default_initial_values (State variables)

These are set by the
@default_parameters and or @default_initial_values

If one is used the reset of the init defaults become the other.
If both are specified the rest become neither

there are two ways that the methods to fill these can be called.

@defaults tag on the class
implementing default_parameters and default_initial_values to call get_map_from_init as AbstractPyNNModel does