Overloading %*%

Spencer Graves wrote:
  

Thanks, Ken. 

      1.  How can I find S4 methods for a given function given class(es) 
of objects?  The 'showMethods' function lists available generics for a 
given function;  "showMethods('%*%')" just produced for me a list of 52 
different signatures for "%*%".  However, I don't know how to find the 
functions with methods for a particular class.  The 'methods' function 
    

Use the 'classes' argument to showMethods, either with or without a 
function as first argument.

  

will produce either S3 methods for a given function or S3 functions for 
a given class.  It would help me if the 'methods' help page included 
"See Also" and "Examples" for S4 classes also, but it doesn't. 

      2.  How can I find source code for S4 methods?  I tried 
    

use includeDef=TRUE with showMethods, or getMethod to get a specific 
method, or, for some fun, selectMethod to find the method to which 
object dispatch occurs. I'm not sure how to dump the method to a file.

Martin

  

"dumpMethods('%*%', 'mmult.R')" and got an apparently empty file of 0 
KB.  Then I tried 'dumpMethod("%*%", c(x="TsparseMatrix", y="ANY"))' and 
got a file with the following: 

setMethod("%*%", structure(c("TsparseMatrix", "ANY"), .Names = c("x", "y")),
NULL
)

      Thanks again for your reply regarding "%*%". 
      Spencer Graves   

knoblauch wrote:
    

Joe Cainey <jcainey <at> gmail.com> writes:
  
      

Is it possible to supply a new method for the %*% operator? 
    
        

clipped
  
      

I've tried to do the same thing with %*%:

"%*%.ad" <- function(a,b)
{
    # further code here
}
However this doesn't work; the new method is never called and the standard
%*% operator is used instead. I've had a look at the documentation and it
appears to be because the %*% operator is not part of the "Math", "Ops",
"Summary" or "Complex" groups. I was wondering if anybody knew of a
work-around for this?
    
        

According to the help page for %*%, it is S4 generic but not S3, so
you might make further progress using S4 methods.

  
      

Thanks,

Joe Cainey

    
        

best,

Ken