[R] [R-pkgs] New version of glmmML

[Douglas Bates]

On 8/21/06, Martin Maechler <maechler at stat.math.ethz.ch> wrote:
> Hi Göran,
>
> >>>>> "GB" == Göran Broström <goran.brostrom at gmail.com>
> >>>>>     on Mon, 21 Aug 2006 11:12:49 +0200 writes:
>
>     GB> A new version, 0.65-1, of glmmML is now on CRAN. It is a major rewrite
>     GB> of the inner structures, so frequent updates (bug fixes) may be
>     GB> expected for some time.
>
>
>     GB> News:
>
>     [............]
>
> Sorry for my lazy question :
>
>    What does the package do that lmer() does not?

Quite a bit, actually.   Göran has the adaptive Gauss-Hermite
quadrature method in glmmML and it is not yet available in lmer.  I
definitely plan to make it available but it has not reached the top of
the "TO DO" list.  Also, as Göran pointed out to me, glmmML works with
the two-column matrix form of the response for a binomial generalized
linear mixed model but the answers from lmer are spurious.  Something
about the weights that I still haven't sorted out.

> I do know that you are a smart person so I know I will get a non-trivial
> answer.  The main reason for this e-mail is really to encourage you (and
> all other 'R-packages' posters) to do the following in each
> posting to R-packages:
>
> 1) [One paragraph:]
>   What does the package do // why do you see a need for the
>   package // what's the bigger context ?
>
> 2) [another paragraph (or two...):]
>   News  etc
>
> Since the R-packages archives are relatively prominently listed
> as part of "News" on the R project website, (and also since I
> encourage ``everyone'' to subscribe to R-packages,)
>
> I think the extra information in "1)" above would be appropriate
> for every posting to R-packages.
>
> Martin Maechler, ETH Zurich
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>