InformationWeek

Secret CIO--2001: An IS Oddity

We may be misguided in our scramble to fix legacy code. How about targeting critically affected business applications instead?

By Herbert W. Lovelace
Issue column appeared: Jan. 27, 1997

Every month, I get together with five of my fellow CIOs in town for lunch at one of the stellar restaurants that grace our fair city. We normally meet at noon and stay until around 2 or so. Our stated purpose is to discuss our jobs in the hope that by sharing experiences, we can continue to be gainfully employed, maybe even earn the modest (our opinion) or ob scene (human resources' opinion) salaries that we are all paid.

The insights we gain, not to mention the good food and beverages we consume, go a long way toward making these gatherings a high point to which we all look forward.

At our last lunch, the conversation turned to the year 2000 problem. As you know, when the calendar hits the year 2000, it could mean disaster for information systems.

Most systems were designed to hold only two digits for the year. So, for example, if you were checking the accrued interest paid for a loan initiated in 1990, the calculation would yield a minus 90 years of debt. While this might be of academic interest to me and my fellow CIOs if we were retired or toiling for our daily wages in a different field--say sanitation engineering--it is of grave concern to us at our present places of employment.

As I listened to my good buddies talk about the heavy dollars they were spending--anywhere from 80 cents to $2 per line to fix the millions of lines of code in their legacy systems--I took another sip of the superb 1991 Beringer Bancroft Ranch Merlot (I recommend it) and decided to enlighten the group with my somewhat controversial proposal. I waited for a pause in the conversation to point out that they were behind the times; they should be worrying about the year 2001 crisis, not the year 2000 problem.

"Ladies and gentlemen," I said, "here is how I see it. We're all going to spend beaucoup dollars to fix the slime we laughingly call our legacy systems. But we are making a basic error in logic. The issue is damage control for the inevitable mistakes we make when we miss patching some critical system. Now, if we don't get on top of this situation, not only will our jobs be in jeopardy, but we'll have even more consultants in our faces. Isn't it enough that we're making them rich explaining the year 2000 problem and the methodologies to fix it?"

I continued, first getting everyone to agree that the major year 2000 cost was finding the problems in the code. T hen I pointed out that it is faster and cheaper to correct the obvious, leave alone what did not affect the business--such as dates on order-entry screens--and build the support capability to quickly fix whatever we miss. I ended by saying that we all knew how spending money to find bugs can quickly reach the point of diminishing returns (unless we're talking about an airline or a hospital), and that we were foolish to let management think otherwise.

The table listened with rapt attention, so I went on to explain that the year 2001 crisis will start with swarms of CEOs saying to CIOs--who have forgotten that it is critically incumbent on them to manage expectations--something like, "You spent $20 million and our mortgage program is wrong!"

I must admit that my argument did not carry the day, even after the second bottle of wine and the excellent créme brûlée for dessert. But then again, these are the same folks who argued so strongly two years ago that everyone would have to conver t to Windows 95 as soon as possible.

Herbert W. Lovelace is CIO at a multibillion-dollar international company. He can be contacted by E-mail at secret@cmp.com . He'll provide real answers (and whimsical comments) to your questions on his Web page, http://techweb.cmp.com/iw/current/secret.htm .

Comments?

http://www.informationweek.com