[R-sig-ME] New alpha version of lme4 code on R-forge

[Douglas Bates]

We have had a bit of grid lock on versions of the Matrix and lme4
packages for the last week or so.  I believe the problem is now
resolved and the development version (0.999375-10) of the lme4 package
on R-forge should be available as a binary package tomorrow.  The
objects produced by mcmcsamp in this version have several kinds of
plots (xyplot, densityplot, qqmath) available for them.  The mcmcsamp
function itself still has the problem of getting stuck at near-zero
variances.  I have been corresponding with Andrew Gelman about methods
that he and co-authors recently described in J. of Computational and
Graphical Statistics to avoid this problem.  These will be implemented
"in the fullness of time", as Bill Venables would say.