[ESS] R help for S4 not opening in help buffer

Martin Maechler maechler at stat.math.ethz.ch
Thu Oct 7 09:47:52 CEST 2010

>>>>> "VN" == Vinh Nguyen <vqnguyen at uci.edu>
>>>>>     on Tue, 5 Oct 2010 08:19:05 -0700 writes:

    VN> On Tue, Oct 5, 2010 at 7:24 AM, James W. MacDonald
    VN> <jmacdon at med.umich.edu> wrote:
    >> Try
    >> ?matrix-class ?show-methods

    VN> Thanks James.  This works for the matrix class.
    VN> However, suppose I load the lme4 package.  Then
    VN> "?show-methods" does not show me documentation about the
    VN> "mer" class show method.  I would have to use
    VN> method?show("mer") to get the right documentation
    VN> (although this also brings me to mer-class since it is
    VN> aliased in the documentation).


    VN> I guess this way of getting documentation for S4 classes
    VN> and methods are not yet implemented in ESS.

That is correct. ESS does not yet deal nicely with the two argument
versions of '?'.
But -- as James MacDonald hinted -- if you know a bit how S4
classes and methods are documented,
you get to the correct help page quickly by trying


and find those  show(<mer>)  methods that are documented with
the class,
or if necessary (not in these cases, and more rarely typically),



Yes, smart dealing with two-argument version of "?" in ESS is something we
(ESS core) know has been on the wish list for a very long time,
and IIRC one reason why we hadn't worked much on solving it,
was that the proposal was that 'R' or the ESS-R interface should
be made more flexible...
One underlying problem is that  "?"  in R  has becomde now
different from  help(.),
and the ESS interface was designed assuming these two were the
same, or rather "?" was a special version of help(), where as
now, these are two different functions, both relying on the same
basic "infrastructure" (lower-level "helper functions", ...).

Patches / Volunteers (both good in R and emacs lisp) 
are very welcome ...


More information about the ESS-help mailing list