[R-SIG-Mac] R on Mac: framework or homebrew?

[Prof Brian Ripley]

On 12/09/2013 17:15, MacQueen, Don wrote:
> I've been using R on OS X probably ever since there was an R on OS X, and
> like you I use it from the command line. In the early days I installed R
> from sources, but quite a few years ago I switched to using the framework
> version. I can't think of any disadvantage I've encountered.

It is harder to have multiple versions of the framework in use at 
once.  And that is an issue now the R release rate is slower: package 
developers probably want both the release version and a current R-devel.

> If I were to not use the framework version, I think I would prefer to
> install directly from sources, not use a package management system. There
> might be more initial work installing various prerequisites (things like
> tcl/tk, png, maybe readline, who knows what all) but it would probably be
> better in the long run.
>
> Some packages, such as a number of spatial packages, require installing
> other libraries such as rgdal, proj, and others. Currently, these are also
> available as frameworks, and installing those packages in a framework R is
> now well-supported. It might be considerably more difficult to install
> such packages in a non-framework R. I don't know how much would be
> involved, so I just suggest it as an aspect to consider.

Almost nothing.  Simon provides binaries for those at 
http://r.research.att.com/libs/ and if you want later ones they compile 
simply from the sources.  in any case, whether R is a framework or not 
does not affect installing packages from sources, as you still need to 
tell the toolchain where to look for the headers and libraries.

On my office iMac have almost all of CRAN installed (more than Simon 
provides as binaries) compiling from source.  The exceptions are those 
10 or so which are too broken to install on OS X and those depending on 
Qt, MySQL (I haven't bothered) or BRugs (i386 Windows/Linux only)


-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595