Judging from the title of this book, you may expect it to be about

• the “year 2000 computing crisis”--the problem that years stored in the common two-digit representation (such as “96” for 1996) can cause bad calculations and bad data, particularly when the year 1999 rolls over to 2000;

• a millennium date conversion project--a project of software maintenance and verification to ensure that such bad calculations and bad data do not occur; and

• a plan for such a project.

If so, you will be disappointed. The authors touch on these topics; they include a 13-page chapter 11 giving advice on a conversion plan. However, the bulk of the book is a description of algorithms for various calculations on dates: converting between various date formats; “aging” (calculating the span of time between two dates); “translation” (adding a span of time to or subtracting a span of time from a date); calculating the day of the week from the date; converting between month-day dates and Julian dates; calculating the date of Easter; and an application of such techniques to build a calendar with dates, days of the week, and US and Canadian holidays.

For each algorithm, the authors discuss its function, the valid domain of the inputs, the return codes, and the results. They give pseudocode in an “IF condition, GO TO label” language suitable for translation into assembly language, and complete listings in IBM Basic Assembly Language (BAL). The BAL programs, plus equivalents in RPG II, are on the diskette included with the book. The COBOL programs on the diskette are not equivalents of the algorithms but examples of how to call the BAL subroutines from a COBOL program.

The book is useful in reminding anyone concerned with date integrity that date calculations and representations need to be handled carefully. (The authors and proofreaders have not been careful enough: there is a typographical error in the very first block of pseudocode, on page 17.)

This book might belong in the library of any group involved in a date maintenance and verification project. On the other hand, it is in no way a good guide to such a project, since one goal of such a project will be to not change the algorithms used in date calculations unless they are actually wrong.

Ironically, the authors threw away a huge lead in devising an effective strategy for dealing with the year 2000 computing crisis, having published an earlier version of this book [1] in 1984, long before the problem became a preoccupation of the trade and popular press and of conference and seminar organizers. To quote the authors: “There has been much procrastination and denial since we last addressed this problem in [1]. A most important resource has been wasted in the interim--time.”