REXX is a structured programming language intended for writing operating system command procedures. It was developed by Mike Cowlishaw around 1980 for use with IBM’s VM/CMS operating system. IBM has selected REXX as its Systems Application Architecture (SAA) procedures language, so REXX programs can be ported from one SAA environment to another. REXX has also been implemented in a number of non-IBM environments, and work is under way to establish the language as an ANSI standard.

This book is a collection of papers about REXX. The authors have varying backgrounds (software developers, academicians, consultants, and users), but all share an intimate knowledge of REXX. As a result, the book contains a wealth of detail about REXX, as well as some information that is not readily accessible elsewhere.

The editors have wisely elected to leave the papers pretty much alone, editing only for consistency. An attempt has been made to impose an organization on the book by grouping the papers into five parts.

Part 1, “General Information,” contains chapters on REXX history, concepts, and programming style. In an apparent attempt at completeness, the editors have included chapters (such as “Maintenance”) that, although useful, contain little REXX-specific information. An interesting glimpse into the future is afforded in the chapter “Object-Oriented REXX,” which describes IBM research into object-oriented extensions for REXX.

Part 2, “Usage,” includes chapters on the use of REXX in the CMS environment to implement applications and macros. These chapters are concerned with such topics as application defaults, routine linkage, argument parsing, error messages, and traps. The chapter “Extending CMS REXX with Function Packages” describes how REXX built-in functions can be augmented or re-implemented, and provides assembler code for implementing a function package.

Part 3, “Platforms,” presents chapters on the implementation and use of REXX in various environments, including CMS (both the interpreter and compiler implementations), MVS, PC-DOS, OS/2, Amiga, Tandem, UNIX, and VAX. Out of place in this part is the chapter “System Product Interpreter Performance Tips”; it contains useful information, but would have been more logically placed in the “General Information” or “Usage” section.

Part 4, “Add-on Products,” contains chapters on products that either enhance REXX (for instance, GDDM-REXX adds graphic capabilities to REXX) or exploit it (for example, DB/REXX uses REXX as a host language).

The title of Part 5, “In Depth,” suggests that it will give more information about REXX implementations, which in fact it does for the S/370, OS/2, and OS/400 interpreter implementations. The other chapters in this part hardly qualify as deep in this sense. The chapters “REXX Education,” “REXX for EXEC 2 Programmers,” and “REXX for CLIST Programmers” have to do with the administrative side of REXX and probably deserve a part to themselves. The final chapter, “Language Evolution and Standards Activities,” is a continuation of the chapter “Procedures Language in SAA” in Part 1. Together these chapters provide useful insights into the language development process within IBM (a process not unlike external standards efforts because of the variety of IBM platforms) and the IBM strategy for REXX extensions.

It is not surprising that all of the authors and editors share great enthusiasm for REXX. What is somewhat surprising is the extent to which REXX is spreading into environments other than CMS (lending credibility to standardization efforts). It is also refreshing to see the willingness of the authors to share the details of their work (including code), which under a less enlightened editorial policy might be construed as intellectual property.

The book will be especially useful to REXX teachers, for additional insights to pass on to students; to REXX implementers, particularly as it helps them understand REXX/operating system interdependencies; and to language standards developers, for insights into the IBM language development process.