Monday, May 27, 2024

The new legacy systems

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

The 1957 Chevrolet Bel Air was a classic, especially the two-door coupe. Tail fins, two-toned paint job, and dual exhausts if you bought the big V-8 engine with the four-barrel carburetor. No seat belts to protect you if a drag race went awry, though. Furthermore, the AM radio wasn’t much to listen to, and the tires, batteries, and other components didn’t last too long, compared to the current versions. Yet, it is undeniable that the car looked good. It brings back pleasant memories of cool summer nights, the energy of youth, and an unbounded optimism about technology and the future. A recent visit to an antique car show that featured a white and turquoise Bel Air (no fuzzy dice hanging from the rear-view mirror) reminded me about another kind of classic that is producing a lot of nostalgia these days–legacy systems.

This chart shows the percentage of respondents to a mail survey who indicated that their ERP system was developed in house.
Source: SG Cowen/Datamation, Summer 1998 User Survey, Networked Computing Servers & PCs

Many IT managers around the world are getting ready to dump their old Big Iron mainframes that run home-grown manufacturing or accounting software. The latest survey of our readers by the SG Cowen Securities Corp., a New York-based investment bank, found that 29% of the responding sites have an in-house ERP system in use this year, but only 19% expect to have an in-house system in operation by the end of 1999. After 20 years of usually faithful service, most of these old Cobol programs are too patched and undocumented to successfully cope with any date after December 31, 1999.

The old system about to be disconnected was the product of countless person-years of systems analysis, programming, debugging, documentation creation, and training. And no doubt the new enterprise resource planning system to replace the old legacy apps will cost your organization a lot more money over the two, three, or more years it takes to implement it.

Don’t opt for the quick and dirty solution

Ever wonder how long that system will last? I’ll bet no company out there is betting that its new ERP package from SAP, Oracle, Baan, PeopleSoft, or J.E. Edwards will last 20 years. The average write-off period may be five to seven years, but I’ll bet the systems being installed this year will still be the core of many organization’s infrastructure in the year 2010. How many of you have planned for that kind of systems lifetime?

Don’t laugh. I know it must seem ridiculous to consider current generation software to have a 20-year lifecycle. How many of today’s vendors will still be around in 10 or 15 years? Hell, how many of today’s Internet companies will be around in 15 months? But your ERP investments may be different. The total cost of implementing these backbone systems has hit $200 million or more for Fortune 100-sized organizations. Procter & Gamble’s CIO, Todd Garrett, recently told me his company’s SAP R/3 project will take seven years to completely install worldwide. I know many other huge companies with R/3 implementation schedules that are more than four years.

In that context, it makes sense to sit back and think about these implementations from a long-term perspective. The next time the implementation team raises an issue that requires a decision that weighs the short-term versus the long-term, don’t opt for the quick and dirty solution. It may be tempting to make your decision based on its impact on your stock options for the year, but in the long term your employer will pay more dearly if you choose the more expedient solution.

Remember that most of our parents and friends outside of IT are incredulous that the relatively innocuous short cut of abbreviating the year field to two digits 20 or 30 years ago will lead to who knows how much chaos over the next 18 months. Not to mention the billions of dollars of wasted money in 1997, 1998, and 1999.

Wouldn’t it be ironic if the new legacy systems we are now implementing will be better than the old systems? Just as today’s cars come equipped with tires guaranteed for 40,000 miles, our new legacy systems may have a longer lifecycle than we expect. If you make sure they’re built to last, maybe they will. //

Larry Marion is the editor of PlugIn Datamation. He has been researching and writing about manufacturing technology for more than 20 years. He can be reached at

Subscribe to Data Insider

Learn the latest news and best practices about data science, big data analytics, artificial intelligence, data security, and more.

Similar articles

Get the Free Newsletter!

Subscribe to Data Insider for top news, trends & analysis

Latest Articles