R help - Mar 2004 - R CMD check warning on predict.systemfit

Hi,

I added a new function "predict.systemfit" to our package
"systemfit" to make
it closer to other packages (e.g. lm). Now "R CMD check" complains
that the
generic function "predict" has only the argument "object",
while our function
"predict.systemfit" has more arguments. However, the function
"predict.lm"
has also more arguments and they are almost the same as in 
"predict.systemfit". Thus, I think that our way to specify 
"predict.systemfit" might be OK in spite of this warning. 
What should I do? Can I ignore this warning? 
What will Kurt answer when we submit it ;-) ?

Best wishes,
Arne

-- 
Arne Henningsen
Department of Agricultural Economics
University of Kiel
Olshausenstr. 40
D-24098 Kiel (Germany)
Tel: +49-431-880 4445
Fax: +49-431-880 1397
ahenningsen at agric-econ.uni-kiel.de
http://www.uni-kiel.de/agrarpol/ahenningsen/

Arne Henningsen wrote:
> Hi,
> 
> I added a new function "predict.systemfit" to our package
"systemfit" to make
> it closer to other packages (e.g. lm). Now "R CMD check"
complains that the
> generic function "predict" has only the argument
"object", while our function
> "predict.systemfit" has more arguments. However, the function
"predict.lm"
> has also more arguments and they are almost the same as in 
> "predict.systemfit". Thus, I think that our way to specify 
> "predict.systemfit" might be OK in spite of this warning. 
> What should I do? Can I ignore this warning? 
> What will Kurt answer when we submit it ;-) ?
> 
> Best wishes,
> Arne
> 
Well, you must use "object" and "..." as arguments, but you
are allowed
to use further arguments. That should not result in a warning.
How does your function looks like exactly?

Uwe Ligges

Hi,

thanks, Uwe and Jari, for your helpful comments! Indeed, I forgot the
"..."
argument. R did not exactly say what was wrong. It just said:

* checking S3 generic/method consistency ... WARNING
predict:
  function(object, ...)
predict.systemfit:
  function(object, data, se.fit, se.pred, interval, level)

and I wrongly assumed that "data, se.fit, se.pred, interval"
shouldn't be
there rather than that "..." was missing.

Thanks again,
Arne

On Tuesday 16 March 2004 11:14, Jari Oksanen wrote:
> Arne,
>
> Are you sure that R warns about extra variables, or does it warn about
> missing parameter "..."?
>
> The syntax of the generic is
>
> predict(object, ...)
>
> and both these should be in your function. You should have "..."
even if
> you do not pass any extra parameters to other functions (been there,
> seen that).
>
> cheers, jari oksanen
>
> On Tue, 2004-03-16 at 11:39, Arne Henningsen wrote:
> > Hi,
> >
> > I added a new function "predict.systemfit" to our package
"systemfit" to
> > make it closer to other packages (e.g. lm). Now "R CMD
check" complains
> > that the generic function "predict" has only the argument
"object", while
> > our function "predict.systemfit" has more arguments.
However, the
> > function "predict.lm" has also more arguments and they are
almost the
> > same as in
> > "predict.systemfit". Thus, I think that our way to specify
> > "predict.systemfit" might be OK in spite of this warning.
> > What should I do? Can I ignore this warning?
> > What will Kurt answer when we submit it ;-) ?
> >
> > Best wishes,
> > Arne
-- 
Arne Henningsen
Department of Agricultural Economics
University of Kiel
Olshausenstr. 40
D-24098 Kiel (Germany)
Tel: +49-431-880 4445
Fax: +49-431-880 1397
ahenningsen at agric-econ.uni-kiel.de
http://www.uni-kiel.de/agrarpol/ahenningsen/