> -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Jeff Gentry > Sent: Thursday, April 21, 2005 9:22 PM > To: Ali - > Cc: r-devel@stat.math.ethz.ch > Subject: RE: [Rd] Overloading methods in R > > > > On Thu, 21 Apr 2005, Ali - wrote: > > I guess the context says you cannot have something like this in S4: > > >setGeneric("testFun", function(...) > > + standardGeneric("testFun")) > > But what is wrong with: > > setGeneric("testFun", function(object, ...))
Naming conflicts, may be the problem. You have that generic in your package, but I might, without know about yours, and I write setGeneric("testFun", function(x, ...)) in my package, and then a third person that we don't know of, is loading both of our packages. Ouch! Indeed, this do happens. I would be ok with the above, if everyone agreed to use 'object' as the first argument - but then you have plot(x, ...) and so on. Also, how should you treat default functions not taking any arguments, e.g. traceback()? Until then I stick with testFun <- function(...) UseMethod("foo") whenever possible. Cheers Henrik > Keeping in mind that 'object' is the actual instantiated > object of the class which you wish to operate on. I guess I > don't see how that aspect is a problem? > > ______________________________________________ > R-devel@stat.math.ethz.ch mailing list > https://stat.ethz.ch/mailman/listinfo/r-devel > > ______________________________________________ R-devel@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-devel