[R] Wald tests and Huberized variances (was: A comment about R:)

Achim Zeileis Achim.Zeileis at wu-wien.ac.at
Thu Jan 5 20:52:07 CET 2006


Peter:

> If R wants to bring in a wider audience, one thing that might help is a
> denser set of cross-references.  For example, perhaps lm's help should
> mention the econometrics view materials as well as other places to look for
> tests and procedures people may want to do w/ lm.  Another thought is that

This is difficult because the core development team has to ensure a
certain stability of the system and you wouldn't start cross-linking to
potentially unstable contributed packages. Furthermore, what is obvious to
you (or me) as further desired functionality for linear models might be
completely counter-intuitive for someone in genomics or biostatistics or
environmetrics or ... and you can't link to all of these without confusing
everybody.

> perhaps the standard R package help should allow people to find
> non-installed but commonly used contributed packages and perhaps their help
> page contents.  A feature that would be very helpful for me is the capacity
> to search all the contents of help files, not just keywords that at times
> seem to miss what I'm trying to find.

This is surely desirable but unfortunately not that simple to implement,
you'll find some discussion in the list archives about this. However,
there are various very helpful search facilites like RSiteSearch().

Best,
Z




More information about the R-help mailing list