Cash for COBOL: A billfold biopsy of the Year 2000 Date Field Problem

-Paul Taylor/Everything 2000

 

The Year 2000 represents a computer maintenance problem of unprecedented magnitude. What's worse, the deadline cannot be moved or missed.

 

$3.6-trillion. That's just a lot of zeros, and they represent the unpaid bill of the Information Technology (IT) industry for what seemed like a good idea at the time.

 

The Year 2000 computer date field problem is rooted in a decades old practice in the IT industry to minimize expensive data storage and data entry time in computer programs by storing the year as two digits (97) instead of four (1997). This becomes a problem when computer systems interpret "00" as "1900" rather than "2000." The problem causes software performing arithmetic operations, comparisons, or sorting on date field to fail or yield incorrect results when working with years beyond 1999.

 

As young programmers pumped out code in the early 1970s to service what are now legacy computer systems, they reasoned that it would all end up on the scrap heap of history long before the new millennium. Some of these same programmers, a little wiser and a little older, now say they made good faith efforts at the time to convince management to go for the eight-digit date field - 2 digits each for century, year, month and date. If the programmers could not envision their code being used a quarter century later, most managers could not see themselves at the same organization in 2000 either. They opted for a six digit date field and saved millions of dollars - at the time.

 

Those savings came at a price that is just now being paid. By one estimate, the worldwide cost of actually fixing the Year 2000 problem has been estimated at $1.5 trillion US. The oft quoted estimate by the Gartner Group, a respected consulting firm that won the ear of Congress in 1996 with an estimate of $600 billion US for a global fix, now contends that the fix is much cheaper than the alternative. Gartner's latest estimates peg the litigation costs alone at over $1-Trillion - citing the potential of companies suing Y2K contractors, citizens suing governments for lost benefits, customers suing corporations for lost goods and services, and everybody else seeking compensation for damages and inconvenience related to inoperable elevators, traffic lights, on board automobile computers and microwave ovens. No matter which way you cut it, the numbers are big and round - suggesting that the only thing of which the estimators are confident is that it will be expensive.

 

Large slices of the conversion price tag will be payable to consultants that take on the risk of converting computer systems for large corporate and government clients. With the stakes so high, Gartner competitor META Group, Inc. might be excused for a dash of hyperbole - "The insidious nature of the Year 2000 date change lies not just in complexity, but also in its scale. As a result, time and resource constraints quickly become the determining factor for success or failure for what are multi-million dollar projects in most enterprises."

 

"Constraints" does not begin to describe the human resources crisis facing organizations as the countdown to 2000 continues. The most dangerous position is that of complacency, thinking an organization is safe because it has a fully staffed Year 2000 project office. Make no mistake: your Year 2000 staff is receiving better offers from other organizations and many of them will leave. Salaries for Year 2000 expertise are rising by as much as 40 percent per year, double the already overheated rate of 20 percent in the IT industry in general. People with the expertise are looking at these offers with a sense of urgency - skills that are at a premium today may not have near the value just 30 short months from now.

 

Some IT professionals have seen the downward swing of this cycle before and are anxious to cash in before it happens again. Case in point, COBOL. By today's standards, it is an arcane computer language and not nearly as sexy as its apparent successors. Yet, COBOL is the underlying language in many of the legacy computers that routinely and faithfully churn out payroll checks, calculate and track mortgages and administer countless government programs.

 

COBOL programmers innovated in developing many of the backbone computer applications in use today but their numbers shrank as development gave way to maintenance. Many programmers moved on to other languages, others moved on to other careers and still others retired.

 

It was this latter group that the former CIO of the State of Oregon set in his sights in developing an admittedly unorthodox approach to finding COBOL expertise for the Year 2000 effort. Curt Pederson tells of visiting every retirement community and nursing home within a reasonable drive of the state capital in search of COBOL programmers. The recruitment drive focused as much on appealing to the retired COBOL corp's sense of public service as augmenting their retirement nest egg. Besides, he mused, it might just be fun to push those legacy systems into the 21st Century.