[Rd] R 2.8.0 qqnorm produces error with object of class zoo?

[Achim Zeileis]

On Wed, 22 Oct 2008, Pfaff, Bernhard Dr. wrote:

> Dear Peter,
>
> many thanks for your swift reply and the enlightenment; much appreciated.
> I have cc'ed Achim as the package maintainer of zoo,

...and I have added Gabor, just in case.

> in case this thread should have slipt his notice. Just curious, Achim, 
> will there be a rank.zoo() function in a future release of your package?

I'm not sure - my first reaction would be "no". With time series, there 
might be confusion between sorting/ordering/ranking with respect to the 
observations vs. with respect to the time index. To clarify that you want 
some operation with respect to the data, you can easily say
   sort(coredata(z))
etc. This is rather evocative and not that much longer to type than 
sort(z). And it would be unclear what the latter should mean exactly.

In the case of rank(z), one could argue that it is meaningful...but 
providing rank() without sort() or order() methods might again be 
confusing. My first impression would be that the cons outweigh the pros.
Z