[R] lme nesting/interaction advice

[Federico Calboli]

On 12 May 2008, at 11:16, Andrew Robinson wrote:
> Well.  I have documentation relevant to nlme that goes back about 10
> years.  I don't know when it was first added to S-plus, but I assume
> that it was about then.  Now, do you think that if the thing that you
> want to do was really bog standard, that noone would have raised a
> fuss or solved it within 10 years?

I'm pretty unpleasant, more so in person, so I'll tell you this. If  
people raised the issue and got the answer I got, I would not be  
surprised if they'd migrated to 'any other stats software' in droves.  
I have no doubt that, given the cryptic and sparse nature of the  
documentation of the issue, most people migrated well before asking -- 
on the grounds most people have a job to do, papers to publish,  
grants to write, kids to pick up from school and pretty little time  
for RTFM and all that sanctimonious attitude.

Once people stop nagging about 'whatever', the reason is because they  
finally got the message things ain't gonna improve, so cut your  
losses and look elsewhere.

Being unpleasant, thick skinned and cheap I will keep nagging and use  
R (the fact I do like it very much might be a factor). But given the  
selection users go through, it will be Vogon time sooner or later ;).

/F

--
Federico C. F. Calboli
Department of Epidemiology and Public Health
Imperial College, St. Mary's Campus
Norfolk Place, London W2 1PG

Tel +44 (0)20 75941602   Fax +44 (0)20 75943193

f.calboli [.a.t] imperial.ac.uk
f.calboli [.a.t] gmail.com