[R] lme (and glmmPQL)

[Pavel Khomski]

Hello!

hier one question about the limits in the number of random effects 
specified by the random part in lme-Call.

could the Call to lme(...) realy fail only becouse of too large 
random-structure part in lme(..., random=....) ?
I really would need many random parameters in the model, to be specified 
for example by the following:


lme(...,
random=list(my.Subject=pdBlocked(list(pdIdent(value=diag(vector_1),form=,nam=),  

                                                                              
pdIdent(value=diag(vector_2),form=,nam=),
                                                                        
      pdIdent(value=diag(vector_3),form=,nam=),
                                                                              
--------------------------------------------------,
                                                                              
pdIdent(value=diag(vector_30),form=,nam=)
                         ))),

...)

Has anybody got such a problem with lme(...) or glmmPQL(...) or  
relative problems?


Thanks a lot in advance.