Also (I haven't really used S, remember, I've just read some
documentation), what makes it different from other "statistical
analysis" software?
--
Jon 'Big J' Alexander, U. of Toronto Comp. Serv.
Toronto, Ontario, CANADA
...!{decvax}!utzoo!utcs!jon
{ihnp4 }
BITNET: jon@utoronto
S is a language and system for analytical computing, data analysis,
graphics, and data management. It provides simple, interactive access
to a wide range of statistical techniques.
The main strengths of S are the simplicity and generality of the
language, the emphasis on interaction and graphics, the
automatically-maintained, self-describing data structures, and the
facilities for the extension of S to new application areas and new
computational methods. Generally, S is more powerful than other
statistical systems in the range of things it can express and in the
facilities for extending this range. It also contains a large number of
specific statistical, graphical, and other analyses.
S is provided in source form and runs under System V, Berkeley 4.2, and
a few older versions of the UNIX system. In general, since S is large,
it is best used only on machines with a large address space (68000
systems, Vaxen, 3B2, 3B20, etc). The critical factor in bringing it up
on a particular machine is usually the way inter-language subroutine
calls are handled. Basically, most of S is written in Ratfor, but the
code that interfaces to the UNIX system is written in C. Hence, it is
important that calls between Fortran and C work well. The C routines
use the calling sequences described in the article "A Portable Fortran
77 Compiler" by S. I. Feldman and P. J. Weinberger, from Volume 2B of
the UNIX Seventh Edition documents.
S can be obtained from:
AT&T Technologies Software Sales
P.O. Box 25000
Greensboro, North Carolina 27420
(800) 828-UNIX
--
Rick Becker alice!rab research!rab