NR backtracking or modify to ensure PSD

- weight & cluster arguments to multigroup function (in case of missing data)

- merge model function

- Simulation with fixed predictors (cond=TRUE) does not work 
  with functional/distribution...

- estimate.multigroup (estimate.lvm) check constraints (default NO?)

- cleanup: estimate.R deriv.R information.R startvalues.R predict.R residuals.R Objective.R

- nested.R

- objective-function - data input.

- auto argument to modelsearch function?

- simulation with fixed predictors

- inklusion af 0-1 variable (Muthen?)

- missing

- check derivatives

- check information

- cumres analytical derivatives

