[ESS] Warning about Emacs/ESS administration and Windows 7 Weirdness

[RICHARD M. HEIBERGER]

> it tells me :
>  "C:/Program Files/q/R-2.11.1/bin"

There are two problems here, both where I consider that R was
incorrectly installed.
1. Where did the "q" come from?  The standard letter there is "R"?
"q" is used as an alternate extension instead of "s" for S+ source files.

2. R should open in the users work area, not inside the R directory structure.
If things are done right, ordinary users should not have write access
to that location.

A sensible place for R to start in Windows 7 is c:/Users/yourname.
On Temple's lab system, R opens in c:/TEMP and students are told to
save everything
to a USB memory stick.  On my machines, for years now, I create a
directory c:/HOME/rmh
and open in there.

>
> Do I just need to manually set ess-rterm-version-paths ?
>
> Also, is there a nice R command that tells me the directory in which I
> can find Rterm.exe?

If your system is non-standard, then according to the doc-string for
ess-find-rterm, you
need to play games with the argument to ess-find-rterm.  That is how I
handled the 32bit and
64bit locations for R-2.11.*.  That might work for R-2.12.* since we
just look down-directory from
the starting location.  I haven't tried this yet, and won't be able to
do so for another two weeks.

This is handled in ess-site.el using ess-find-rterm which I designed
based on ess-r-versions-create
that you wrote.  I think you modified it further.  ess-find-rterm
finds all versions of R that are installed in
the standard places.  That was the cause of the fuss when Windows64
changed the definition of standard,
and is the reason I need to make modifications for R-2.12.0 which puts
both 32bit and 64bit R in a new
location, one step down in the directory structure.



> ... and then I'm thinking of whether we can get R to include a default
> function, say ess.find.rterm() that could automatically generate the
> lisp code that a user would need to stick in their ~/.emacs so that that
> version of R is found by a fresh emacs.

That is an interesting idea.  It would be easy, but I am not sure it is needed.


>
> Stephen
>