R-beta: Splus vs R

Paul Gilbert (la-jassine@aix.pacwan.net)
Fri, 02 May 1997 17:19:04 -0300

Date: Fri, 02 May 1997 17:19:04 -0300
To: R-help@stat.math.ethz.ch
From: Paul Gilbert <la-jassine@aix.pacwan.net>
Subject: R-beta: Splus vs R

> If I were Mathsoft I would be less than pleased
> at the development of R and would try to stop it if I could. 

I would not assume that Mathsoft must have such a negative view towards R.
The reasons are roughly summarized by the analogy that a smaller part of a
big pie is often better than a bigger part of a small pie.  

Mathsoft's limited success in addressing the student market has already been
mentioned, and the longer term implication of this would be fewer S users
and programmers. Companies like SPSS have addressed this by offering very
good prices for student versions which lack features  needed by almost
everyone with any money. There are difficulties using this approach with a
programming language like S, especially when there is such a large base of
free application programs available at Statlib.

The other end of the market has not been mentioned. Large organizations,
with money that makes Mathsoft's profit margins look more interesting,  let
end users like statisticians use S. Their IT departments often tend to think
of S as an application program and any S code as application macros. In a
few cases the IT departments have adopted S as a programming language. Often
(possibly even usually) these departments have a "second source" policy.
That means a programming language cannot be used unless there is more than
one source for compilers and any other programs necessary to actually use
any code developed. In the past the need for a second source may have been
met by the fact that S (not Splus) was available from Bell Labs. Since that
is no longer the case, the possibility for IT organizations to adopt S has
been severely limited, at least until a completely functional version of R
is  available. The possibility the Mathsoft would actually lose these
customers to a free software product  is almost non-existent. Generally the
support that Mathsoft offers is a real bargain compared to the cost for an
organization to provide similar support internally. (Some  would also say
that IT organizations need someone to blame when things do not work, and
blaming a source that provides free software with no warrantee is not good
enough.) Usually, the fact that R may be faster than S would not be a factor
(as it is more likely to be for academic researchers). There are lots of
applications and operating systems around to illustrate that slow speed does
not prohibit commercial success. In the long term, a broader adoption of S
by IT organizations will insure continued development and the viability of
S, Splus, Mathsoft, and R.

I hope that Mathsoft takes this broader view toward R development as I think
R is important for the whole community, even for Splus users who may never
consider the possibility of using R.

Paul Gilbert

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@stat.math.ethz.ch