R devel - Jan 2004 - non-generic functions

Hi,

I'm writing some code that will eventually see its way to a package (I
hope). I've noticed that some functions are not set up using generic
methods (while similar functions are). For example, mean is a generic
method but var is not. Other operations that I've come across that could
benefit from being generic are: qqplot, cor, cov, optimize/optimise.

For example, if one were developing a control charting package, it be great
to simply use sd(object), where object is of course the appropriate class,
to get the standard deviation of the process as estimated be the charting
method. Similarly, a DOE package might include optimize(object) to find the
optimal operating conditions based on the data and specifications in
object.

Just a thought,

Max



**********************************************************************
This message is intended only for the designated recipient(s...{{dropped}}

If you use S4 classes/methods you can "create" generics by just
calling

setGeneric("sd")

for example, and the original function becomes the default method.  Then 
you can define appropriate methods for your objects.

-roger

Max_Kuhn@bd.com wrote:
> Hi,
> 
> I'm writing some code that will eventually see its way to a package (I
> hope). I've noticed that some functions are not set up using generic
> methods (while similar functions are). For example, mean is a generic
> method but var is not. Other operations that I've come across that
could
> benefit from being generic are: qqplot, cor, cov, optimize/optimise.
> 
> For example, if one were developing a control charting package, it be great
> to simply use sd(object), where object is of course the appropriate class,
> to get the standard deviation of the process as estimated be the charting
> method. Similarly, a DOE package might include optimize(object) to find the
> optimal operating conditions based on the data and specifications in
> object.
> 
> Just a thought,
> 
> Max
> 
> 
> 
> **********************************************************************
> This message is intended only for the designated recipient(s...{{dropped}}
> 
> ______________________________________________
> R-devel@stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-devel
>