AWM Book Review

Time Bomb 2000

Edward Yourdon and Jennifer Yourdon, Prentice-Hall, Upper Saddle River, NJ 1998. xxv+416. ISBN 0-13-095284-2 (paper), $19.95.

From: AWM Newsletter, May/June 1998.

Reviewed by: Marge Murray, Book Review Editor, Department of Mathematics, Virginia Tech, Blacksburg, VA 24061-0123; email: murray@calvin.math.vt.edu.

This column is usually devoted to the latest books on gender, science, mathematics, and education. The present review is clearly a departure from the usual fare. I have chosen to review this book, in this issue, because I firmly believe that the year 2000 problem is a serious one, with which every member of AWM should be familiar, and for which every member of AWM should be prepared. Time Bomb 2000 offers common-sense risk assessment and balanced, reasonable advice on how to prepare for the most likely year 2000-related disruptions. It is by far the best book on the year 2000 problem for the layperson that I have yet seen.

What is the year 2000 problem? In a nutshell: computers and microprocessors that use six-digit dates (e.g., dates in the format MM/DD/YY, with two digits each devoted to the month, day, and year) will have difficulty recognizing and processing date information when the year 2000 rolls around.

Why? Well, suppose you use your Visa card to pay for the holiday gifts you buy during the month of December, 1999. And suppose the closing date of your monthly Visa statement is January 3, 2000. The bank's computer will look at the $100 in purchases you bought on Christmas Eve (12/24/99) and subtract the date of purchase from the current date (1/3/00). Now, if the bank hasn't solved its year 2000 problem, one of two things will happen. Either (a) the computer will spit out an error message, because the date of purchase is later than the current date -- and the bank won't know how to bill you; or, (b) taking the absolute value of the difference in dates (as banking programs sometimes do), the computer will determine that your bill is over 99 years overdue. Assuming continuous compounding of interest at an annual rate of 16%, what will the bill for those $100 of last-minute Christmas gifts amount to? (Try giving this problem to your sophomore calculus/ODE students!)

Now, on first glance, this seems like a ridiculous problem --- and one that should be very easy to solve. The bank's data processing department simply needs to re-write its programs so that dates are represented in eight-digit format (MM/DD/YYYY), and all will be well. Indeed, my very own bank has sent me a comforting notice in my most recent bank statement saying that they are "committed to solving " their year 2000 problem -- not to worry!

But the problem is in fact more serious than it would at first appear. Nowadays, computer memory is incredibly cheap, and getting cheaper all the time; but in the early days of computing, memory was precious, and dates and date-fields in computer programs were assigned six digits in order to conserve memory. In fact, in 1970, the National Bureau of Standards (now the National Institute of Standards and Technology) made six-digit dates a standard for federal computer systems. Now, difficult as this may be to believe, many businesses -- federal agencies, banks, brokerage houses, manufacturers, etc., etc. -- are still using six-digit dates; what's more, many of them are still using computer programs written in the sixties and seventies to do their date-sensitive computing. Six-digit dates are embedded in hundreds of millions of lines of computer code -- indeed, six-digit dates are ubiquitous in financial transactions all over the world.

Even if my own bank fixes all of its year 2000-related programming problems, its computers must communicate with computers at hundreds of other banks and businesses, not to mention the computers of the Federal Reserve System. The financial environment in which we live and work depends upon a complex web of interconnected, intercommunicating computers. A year 2000-related failure in one part of the financial system is likely to create a ripple effect of unintended consequences that could spell disaster even for those who thought they had their year 2000 problems solved.

Although computer professionals have recognized the problem for many years, no one really expected that so many old computer programs would still be running -- or that computer programs would still be using six-digit dates --- as late as 1998. But it is only in the last three or four years that the data processing departments of businesses large and small have begun to pay attention, and to devote resources to the massive reprogramming efforts needed to prevent year 2000-related business disasters. And over the past three or four years a whole host of technical articles, books, and websites have cropped up offering assistance to computer programmers and data processing managers in their effort to get these problems fixed on time.

On March 19, 1998, The New York Times reported that "with only 653 days to go, it is too late for Federal agencies to fix all of the year 2000 bugs in their critical computer systems, and some Government functions are likely to be disrupted." Experts in the private sector estimate that at least 15% of all year 2000 bugs will go unrepaired before the deadline of January 1, 2000. What seems certain at this point is that our increasingly computer-dependent lives and lifestyles are going to be seriously disrupted in the year 2000. The question is: how serious is the disruption likely to be, and what can we do to prepare for it?

This and related questions are addressed in the book Time Bomb 2000, written by the father-daughter team of Edward and Jennifer Yourdon. Despite the sensational-sounding title, the book offers cogent, common-sense advice on how to prepare for a possible breakdown in the computing infrastructure of our lives. The initial working title for the book was the less alarming Fallback, since the book offers practical advice on what sorts of strategies one can "fall back on" in the event of a two-day, one-month, one-year, or ten-year disruption in the everyday, computer-regulated functioning of postmodern life.

The book is organized into topical chapters, which assess the effects of uncorrected year 2000 bugs in such areas as jobs, utilities, transportation, banking and finance, food distribution, health and medicine, education, telephone service, and so forth. Perhaps the most eye-opening chapter deals with the year 2000 problem in "embedded systems" -- the tiny microprocessors found in automobiles, microwave ovens, VCRs, heart pacemakers and defibrillators, industrial process control systems, and indeed, every piece of electronic equipment which contains a microchip. Yes, computer engineers have replicated the six-digit date format in a good many of these chips -- the critical question is, how many? The Yourdons write:

[O]nly a small percentage of these chips are likely to be "year-sensitive," and only a small percentage could be described as "mission-critical" (in the sense that a failure could cause severe economic consequences and/or loss of life) -- but even a small percentage of a small percentage can be a large number when we start with a population of 25 billion chips. The vast quantity of these embedded systems is one of the key issues that you need to keep in mind. There are simply not enough programmers, not enough repair technicians, and not enough time to repair and/or replace them all; even if only one tenth of one percent of the 25 billion chips are year 2000-defective, that still leaves us with 25 million repair jobs. (p. 285)

This excerpt offers a glimpse into the size and scope of just one aspect of the year 2000 problem.

I am not normally a devotee of doomsday literature -- and despite its scary title, Time Bomb 2000 is not a doomsday book. It is a practical, balanced guide to a frightening problem which, for all its simplicity, is absolutely stunning in its complexity. In their concluding remarks, the Yourdons write:

We are optimists at heart, and we agree with the late Mary McCarthy's asssessment: "The happy ending is our national belief." But we are also deeply concerned about the impact of year 2000 software problems in every aspect of our lives. The more we've investigated the situation during the course of preparing this book, the more worried we have become. We would like to believe that actions by computer technicians, business executives, and government leaders during 1998 and 1999 will give us cause for optimism -- but we are reminded of Oscar Wilde's observation in The Picture of Dorian Gray: "The basis of optimism is sheer terror." In the final analysis, we believe that it's better to be terrified now and take appropriate actions, even if it turns out that the year 2000 problems are no worse than a few mosquito bites. The alternative -- being complacent now and facing the possibility of severe year 2000 problems without any fallback plans -- could turn out to be the equivalent of a fatal rattlesnake bite. Consequently, we're making our fallback plans now -- and we hope this book has convinced you to begin making yours. (pp. 347-348)

I urge you to go out and buy this book.