[ADMB Users] The current state of glmmADMB

[Arni Magnusson]

The R package 'glmmADMB' is recovering from a somewhat confusing state, 
which is worth explaining.

The background is that Hans Skaug, Dave Fournier, and Anders Nielsen wrote 
glmmADMB back in 2005 and it has been available at 
http://otter-rsch.com/admbre/examples/glmmadmb/glmmADMB.html. That page 
now has the legacy version of glmmADMB 0.3-0 (Oct 2006), and some useful 
info.

When R 2.10 came out a year ago, old packages like glmmADMB 0.3-0 started 
giving warnings in Windows, saying that they needed to be rebuilt, because 
package infrastructure had changed between R 2.9 and 2.10. The 
infrastructure changed again this month, between R 2.11 and 2.12, but 
that's another story.

This summer, I suggested to Hans that glmmADMB should be moved to R-Forge 
or CRAN, so that users can install glmmADMB for the current version of R, 
and to make it more visible to potential users. Sure, he said, go ahead.

The official home of glmmADMB is now on R-Forge 
(http://glmmadmb.r-forge.r-project.org). Ben Bolker and I have joined the 
original authors and made some internal improvements to conform to modern 
R package conventions. glmmADMB will not be available from CRAN, because 
of restrictions regarding binary executable files. That's fine; R-Forge is 
an excellent server for R packages, that makes it easy for multiple 
authors to collaborate. The current version of glmmADMB is 0.4-1.

---

The only problem is that the R-Forge package server is currently in a 
shell shock, probably due to the aforementioned package infrastructure 
changes between R 2.11 and R 2.12. The server hasn't updated any binary 
packages http://r-forge.r-project.org/bin/windows/contrib/latest/ since 13 
October. Therefore, the normal way of installing glmmADMB,

   install.packages("glmmADMB", repos="http://r-forge.r-project.org")

only works in Linux right now. To install glmmADMB 0.4-1 in Windows, you 
need to specify type="source":

   install.packages("glmmADMB", repos="http://r-forge.r-project.org",
                    type="source)

There are no statistical differences between glmmADMB 0.3-0 and 0.4-1, so 
they give the same model fit. We just removed some unnecessary warnings, 
improved the help pages, and refactored the R functions. The most 
important change is the move to R-Forge, and I wanted to tell you how to 
stroke it the right way with type="source", to bring everyone on the same 
0.4-1 page.

Arni