[Rd] future update to loess

[Benjamin Tyner]

Background: I'm a student of Prof. Cleveland at Purdue University.
Eventually, we'd like to release a new version of the loess routine in R.
For starters, this implementation would have support for local polynomial
degree 3, better control over the number of cells in the KD tree, and
perhaps a better solution in higher predictor dimension.

I see that Prof. Ripley was responsible for originally porting loess to R
as part of the modreg library, which has been absorbed into the stardard
stats library. My first question is, what would be the preferred way to
package this update? As an updated version of modreg, or as a new standalone
implementation of loess? Or something else entirely?

Second, I'd like to hear what additional features or modifications other
loess users would like to see incorporated into this release.

Thanks,
Ben