RE^2: [Rd] New generic functions in "R base" {was `truncate'}

Kurt Hornik
Fri, 25 May 2001 19:34:36 +0200

>>>>> Byron Ellis writes:

> I think the proposal was for something along the lines of Perl's CPAN
> module (or ActiveState's PPM for that matter)---sure, you have to
> update indices periodically, but whatever. Perhaps some sort HTTP
> based system for performing full text queries (since the database for
> that sort of thing will undoubtably be large) would be in order, but
> the spirit of the idea is certainly a good one---especially for new
> users (students, etc) who don't know the system like the back of their
> hand. 'sides, kids these days aren't used to things like manpages and
> apropos. ;-)

> ...

Sorry for entering this a bit late.  The CRAN maintainers have actually
decided implementing something along the above lines a long time ago.
Unfortunately, it simply has not happened thus far.  In a way, this
would be a good project for a computer science student ...

In a nutshell, I see two approaches.  One amounts to having a web front
end for which Brian kindly pointed out.  I.e., CRAN would
have a prebuilt help db for all packages it provides, and a search
mechanism patterned after [you do not necessarily need an
R process for the search].  Full-text search fanatics might prefer us to
simply `install' all HTML files of the CRAN packages to some place and
provide full-text indexing and searching of that.

[To those who haven't read from me about this: full-text index/search
alone does not solve the problem.  You want fuzzy matching a la glimpse
along with that, otherwise e.g. spelling differences cannot be

As we hope that R 1.3.0 will support multiple package repositories, the
question *where* to search will become more delicated.  For example, the
packages by Jim Lindsey are not distributed via CRAN ...

If there are any volunteers for implementing this, pls let us know.


r-devel mailing list -- Read
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: