[ADMB Users] Integrating R and ADMB to generalize TPL files

[dave fournier]

But that is not right.  The model is to assume that the
random effects in different time periods are independent and that their
variances are different.  You can do this by figuring out what the
correct covariance is for the current design matrix or by
writing the model properly which is trivial to do in ADMB.  As usual it
is the R interface and these silly design matrices which introduces the 
problem.