[R] Re: the R package - relation to R project?
casgrain at magellan.umontreal.ca
Thu May 17 05:55:39 CEST 2001
Hello, I'm the head developer of the R Package, so I'd like to clarify
some things :-)
> I stumbled across this link and was wondering if this has any relation
> the R project.
Aside from using the 18th letter of the alphabet, none. "The R Package"
predates the "R Project" by several years. The word "package" stems from
the fact that it was, in the beginning, an eclectic collection of
multivariate analysis programs.
> However, even Pierre Legendre's R is freely distributed (although
> I am not sure they are very careful about licensing of all components
> they use).
When I get around to finish cleaning it up and bringing everything to
version 4, we're thinking about releasing it as open source. However,
it's all written in Pascal, and Pascal compilers are not very common in
open-source projects. Besides, IIRC, all the algorithms are described in
Legendre and Legendre's "Numerical Ecology" textbook (1998).
> This is a project Pierre Legendre and its staff from Montreal University
> started 5 or 6 years ago. Legendre's R package has nothing to do with
> R! It
> is a collection of routines for numerical ecology for the Macintosh (and
> everything can be nicely done in R -the CRAN R, I mean-). This is
> because both projects have almost the same name. What to do? Obviously,
> of these projects will change its name!
5 or 6 years? Try 23! It actually started in 1978. I'm the third
programmer on it. It's also not a collection of routines, but of
compiled applications (currently only for MacOS, but I'm working on a
Windows port). If one were to follow rules of publishing, we'd have the
original ownership since we published it first ;-)
> Why not? Perhaps 'our R' should change its name to 'R-plus'? :)
Since R is related to S+, I believe that makes sense! That, or a
recursive acronym à la gnu: RinS (R is not S plus)...
> I remember NAME, an European seminar on Numerical Analysis in Marine
> where Pierre Legendre proudly introduced its R package to us. We were
> wondering why this name, R, for such a software... and we finally
> that it should be because it adopts an R strategy.
Nice guess :-), but here's the real reason (from our user's guide):
"The name of the package, “R”, comes from our 1978 work on PDP-10
equipment, where R (for Run) was the command to start a program. In the
CDC implementation, we kept “R” as the name of the main calling program
written in CCL, which is the equivalent of a large IBM EXEC file, thus
mimicking the PDP way of calling the programs. This is why “R” has
gradually shifted to become the pleasantly short name of the package."
> Anyway, if R is mysterious as a name for Legendre's R package... it is
> clearer for "CRAN-R" because it is "not unlike S"! I suppose that if
> somebody wants to make another software "not unlike R", he should call
Funny, I'm actually working on a program whose code-name is "The Q
Package" since it is partly derived from the R Package, and will replace
another (commercial) product whose name begins with the letter Q...
Best wishes to all,
This message was brought to you by the letter R
r-help 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-help-request at stat.math.ethz.ch
More information about the R-help