My hyperSpec
package provides functions to work with spectoscopic data sets. I provide hyperSpec-methods for quite a number of functions coming from other packages. However, some of these other packages, I have listed as Suggests:
as I do not want to force users to install something like 20 other packages (counting dependencies) if they may never use the specific functionality.
From the coding point of view, that works nicely. However, I'm at a loss how to deal with that in the .Rd file - but I'm sure it is a rather common problem.
The usual \code{\link[otherpkg]{fun}}
will not reliably work, as there is no guarantee that otherpkg is available. If otherpkg is not installed, then the link is broken. That will be detected by R CMD check
e.g. on Winbuilder and r-forge. Which means: the package won't pass CRAN checks.
Which packages would be a good example to look at?
The only hint (though no solution) I found so far: https://stat.ethz.ch/pipermail/r-devel/2009-November/055425.html (but there was no answer to that email).
otherpkg
is installed, otherwise the link is broken. Ifotherpkg
is inSuggests:
but notDepends:
, I cannot rely on that link. I clarified the question. – Hoopoe