[R] namespace question

[Prof Brian Ripley]

On Mon, 1 May 2006, Erich Neuwirth wrote:

> I need to modify the function "loess.smooth" from package stats
> to accept an additional weight parameter.
> I found the places where I need to change things,
> but there is a problem remaining (in R 2.2.1).
> loess.smooth contains the call
>
>    fit <- simpleLoess(y, x, w, span, degree, FALSE, FALSE,
> 		       normalize=FALSE, "none", "interpolate",
> 		       control$cell, iterations, control$trace.hat)
>
> Since simpleLoess is not exported from stats,
> I need to change this to
>
>    fit <- stats:::simpleLoess(y, x, w, span, degree, FALSE, FALSE,
>
>
> Furthermore, simpleLoess does
>
>    z <- .C(R_loess_ifit,
>
> Changing this to
>
>    z <- .C(stats:::R_loess_ifit,
>
>
> solves the problem in R 2.3.0, but not in R 2.2.1
> Is there an easy solution for 2.2.1 also?

The call in 2.2.1 was .C("loess_ifit" ..., and does not need changing.
The form used in 2.2.1 also works in 2.3.0 (just a little less 
efficiently).

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595