R-alpha: R 0.50.a1 problems with lm/glm methods

Kurt Hornik Kurt.Hornik@ci.tuwien.ac.at
Mon, 28 Jul 1997 15:23:48 +0200

Thanks for fixing formula() and update().

Here are some open problems.

* effects:

** There is no help available for effects()

** The lm() help page has effects(lm.obj), but from the code it is clear
that the one-arg version cannot work.

Shouldn't it simply return z$effects?

** The two-arg version only works for factors (?) ... otherwise,
	z$model.frame[factors & pattern != 0]
gives an error.  Can someone explain to me what this is doing?

* plot:

There should be a plot.lm() function which reports that useful
diagnostic plots have not been implemented yet.

Perhaps we should do something like

	unimplemented <- function () {
		stop("Not implemented yet ... sorry.")

	plot.lm <- function (...) unimplemented()
	plot.glm <- function (...) unimplemented()

and similarly for other stuff which is known to be missing?  (This would
also help improving the TASK handling :-))

* predict:

** There is no help available for predict()

** The one-arg version of predict.default() cannot work.  For objects
of class "lm", it should simply return the fitted values, right?

** The two-arg version of predict.default() cannot work because no
fitted object has a "covmat" component.  Anyway, shouldn't the general
form be something like
	predict(object, newdata, ...)

r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch