S (programming language)

S is a special kind of programming language made mainly by John Chambers and a team at Bell Laboratories. They wanted to help people turn their ideas into working computer programs fast and accurately. This language was very useful for doing statistics and data analysis.

Many teachers and researchers used S for their work. But today, a new language called the R language has taken its place. R is partly compatible with S and is part of the GNU free software project. There was also a version called S-PLUS, which was sold by TIBCO Software for business use.

History

"Old S"

S is a statistical computing language created at Bell Laboratories between 1975 and 1976. Before S, people used Fortran subroutines for statistical work, but S offered a more interactive way to work with data, inspired by ideas from John Tukey. Important features from the start included interactive graphics and easy-to-find help for functions.

John Chambers and Trevor Hastie led the development, with help from Richard Becker, Allan Wilks, and William Cleveland. Chambers is considered the main creator and won an award for his work on S. The first version of S appeared in 1976 and ran on the GCOS operating system. It didn’t have a name at first, but eventually, it was called “S” after the word “statistics.” In 1979, S was moved to the UNIX platform, making it easier to use. By 1980, S was shared outside Bell Laboratories, and in 1981, the source code became available. It became popular among academic researchers. Two important books about S were published in 1984.

"New S"

In 1988, a new version called S-PLUS was released, later sold to TIBCO Software. This version changed many parts of S, such as switching from macros to functions. A book called The New S Language explained these changes. Users found it hard to adjust because their old code needed rewriting. Other changes included better graphics and updating many inner functions. The new version of S was similar to later versions of S-PLUS and R.

In 1991, a book called Statistical Models in S introduced new ways to define statistical models and changed how objects and methods worked.

S4

The latest version, S4, came out in 1998. It added advanced features for organizing data, with clear rules for each type of data and how they connect to each other. It also allowed functions to change behavior based on many inputs, not just one.