****
We should consider making key entities objects and writing accessor functions to them,
like modelGet, modelSet ...

function rmModelGet

switch lower(param)
    case {'x0','xeccentricity'}
        val = mdl.x0;
    case ''
    case ''
    otherwise
end


return;

%------
function mdl = rmModelSet(mdl,param,val,varargin)
%
%
%
%
%

switch lower(param)
    case {'x0','xeccentricity'}
        mdl.x0 = val;
    case ''
    case ''
    otherwise
end

return;

% ---
Code will look like this
eccentricity = mdlGet(model{1},'ecc');

% -----
Don't set anything that could conflict.  So set, say, x0,y0 and then
derive eccentricity

rmModelGet(model,'eccentricity')

rmModelGet(model,'x0').^2 ....

****
Find a way to organize the output files.  THere are so many.  We should build a
better directory structure for them?  Or should make a new data type?

****
I think we should break out some of the routines.  

****
Visualization of the model information.
Should we project the model parameters back onto visual field reprsentations?
For example, select a point and show the position and shape of the
estimated RF on an image of the visual field.

