I put out an email last night with output from my package check. Many
thanks to Uwe and Liviu who got back to me quickly about how to fix some of
the errors. I realize though, that I had some additional questions /
confusion points in terms of documenting packages.
1) I seem to understand how to document a straightforward function, but am
having trouble figuring out how to document methods functions when creating
new S4 objects. For instance, I am trying to document the following
function:
setMethod("tail", "NFIRS", function(x, n=5)
tail(as.data.frame(x), n=n))
So far, I have the following for my .Rd file (built automatically by
package.skeleton)
\name{tail-methods}
\docType{methods}
\alias{tail-methods}
\alias{tail,ANY-method}
\alias{tail,NFIRS-method}
\title{ ~~ Methods for Function tail ~~}
\description{
~~ Methods for function \code{tail} ~~
}
\section{Methods}{
\describe{
\item{x = "ANY"}{ ~~describe this method here }
\item{x = "NFIRS"}{ ~~describe this method here }
}}
\keyword{methods}
\keyword{ ~~ other possible keyword(s)}
Since the tail method is already defined for R and my function works the
exact same, do I still need to make a man file for this function? Do I just
copy the information from the R tail function (i.e. Title: "Return the
First or Last Part of an Object"). What do I put down under describe (the
"items" don't really make sense to me since my tail function
requires two
things 1) an object of type "NFIRS" and 2) n, the number of rows to
display
-- I don't understand what x="ANY" and x="NFIRS" means)?
Do I need to add
any keywords, or can I just erase the "other possible keywords" line?
2) Do all functions in a package need to be documented even if the user
will never use them. I have lots of little support functions that the user
should never really use. Just wondering if all packages submitted to CRAN
are required to have every last function documented.
Thanks for your help.
--Markus
[[alternative HTML version deleted]]
