twinned case must be considered

hyperspy/model.py

@@ -1191,6 +1191,12 @@ def fit(self, fitter="leastsq", method='ls', grad=False,
                     raise AttributeError(not_linear_error + str(nonlinear))
                 # INCLUDE CHECK that there is at least 1 COMPONENT - if none,
                 # raise error
+
+                # TODO: May need to add all components that are linearly twinned with
+                # a component in order to scale them correctly. Should try and see what would 
+                # happen if I had three gaussians ABC, where B is twinned and follows A, but B and C
+                # are in similar positions. Will C be overfit in the position where B would normally sit?
+                #  
                 self._check_and_set_linear_parameters_not_zero()
                 self._set_p0()
                 number_of_free_parameters = len(self.p0)