[Home Page] - [Reviews Main]

The Design of OS/2 by Deitel & M S Kogan Recommended

ISBN: 0-201-54889-5       Publisher: Addison-Wesley       Pages: 388pp       Price: £32-25.

Categories:   os/2

Reviewed by Neil Bowler in C Vu 4-4 (May 1992)

The details in the book come from "the horse's mouth". One author is the lead architect of OS/2 2.0. This leads to some statements being biased towards IBM. It does, however, give an insight into some of IBM's strategies such as Systems Application Architecture (SAA), and where OS/2 fits into the scheme of things.

Although the book is based on OS/2 2.0 it gives a full description of the internals of OS/2 1.x, and details differences between the two main versions of the operating system. The book makes some comparisons with Unix, but this is usually in the context of "OS/2 does it this way, UNIX does it that way, therefore OS/2 is much better".

The book is well structured, and fairly readable, but the nature of the book makes it heavy-going at times. Each chapter is contains a brief summary and a list of the terminology used (which is quite large for most chapters). A glossary at the end of the book might have been more useful. Each chapter concludes with a section of exercises based on that chapter. This seemed odd, perhaps the authors are intending the book as a course text-book.

The first 80 pages concentrate on historical background, microprocessor architectures (80x86 and RISC) and hardware architectures. The book then gives a brief description of the essential workings of MSDOS. Though interesting I question its inclusion in a mid-length, expensive book. All of this subject matter is available in other books many listed in the good 8 page bibliography.

The remaining chapters describe the various aspects of OS/2, starting with the base operating system; multitasking, memory management, interprocess comms, and I/O management. It then goes on to describe how compatibility for MSDOS, OS/2 1.x and Windows is produced. Finally, a description of the various communications protocols supported by OS/2 and a description of the issues that the authors believe are important to the future of OS/2; open systems, SAA etc.

The book gives a good technical description of the fundamentals of OS/2 which should be invaluable to any programmer wishing to make full use of its capabilities. It also gives a good indication of what OS/2 2.0 is like to people unable to obtain a Beta-test copy of it (although I would certainly not buy it on the strength of that alone). If you can not justify the thirty or so pounds the book costs, try to at least borrow a copy to read. This book contains many useful details that I am sure will not be available in any of the manuals IBM will release.