Old Technology Never Dies

How many times have we heard of the imminent demise of COBOL? Or perhaps the imminent murder would be a better question, because COBOL's end is always going to be delivered by some other technology.

Over the past 30 years, COBOL was going to meet its maker at the hands of fourth-generation languages, program generators, CASE (computer-aided software engineering), object-oriented programming languages, integrated development environments, and component-based development.

But COBOL lives on. To this day, there are more COBOL application systems in production than there are systems written in any other language.

COBOL's would-be assassins, however, didn't simply run into the COBOL monolith and drop dead in their tracks. They too survived and today we have plenty of production systems developed in those technologies as well.

This pattern is actually common: A technology becomes king of the hill. Subsequent technologies are supposed to depose the king, but today the original technology and most of its erstwhile deposers perform side by side. Examples that come to mind include mainframes and client-server, desktop PCs and thin clients, and traditional business practices and e-business.

Those who bemoan the consequent complexity generally hold out some kind of homogeneous technology architecture as a holy grail. They resent the apparent waste of supporting multiple technologies that provide similar functionality.

One example of this often occurs when reviewing hardware budgets. "Why should we pay maintenance fees for the mainframe, AS/400s, and our UNIX servers when we could just run everything on UNIX?" is a typical question. On the software side it often looks like this: "Nobody's going to be writing Visual Basic in another two years, so we should convert everything to Java now."

Regardless of the form it takes, the argument is always based on the assumption that we should reduce technological complexity as much as possible. If we had no other goals, this would be laudable. The problem is that business always has other goals - goals that are more important than technical homogeneity.

The consequence is that old technologies never die.

The resulting heterogeneity is a natural byproduct of the way new technologies are used: They are used to create new systems and extend the functionality of existing systems. They are generally not, in contrast to popular belief, used to rewrite existing systems.

This is only natural from a business perspective. If you have systems that are working well enough, there is no reason to rewrite them if the only benefit is technology homogeneity. Scarce business resources are better spent producing greater functionality to enable the business to do more than it can now.

This carries several implications for IT.

First, forget homogeneity. Plan on running application systems until they are absolutely ragged. Then and only then will the business be willing to devote resources to replacing them. This means that you must plan to support the technologies these application systems depend on long past their day in the sun.

Secondly, and corollary to the previous point, before adopting a technology be prepared to live with it for a long time. This means being willing to pay apparently redundant maintenance costs and ensuring that at least some members of your staff remain conversant with the technology until it is retired.

Another implication, one that will help IT stay aligned with the business, is: When a new technology comes along, look for ways to use it that extend or enhance existing functionality.

It is always functionality that is important to the business, not the technology behind it. Even when a new technology is the sexiest thing in the world, there is no way that the technology should be adopted just because it's the latest and greatest.

Only when business needs require functionality that the technology is well suited to deliver should the technology be adopted. And then only to provide that functionality, not to convert existing applications to the new technology.

Technology heterogeneity is a natural result of a successful business. One consequence of this is the fact that old technologies never die - and that's just the way it is.

Chris Pickering is president of Systems Development Inc., an IT research and consulting firm. He also is a senior consultant for the Cutter Consortium. His latest industry survey, Strategic Trends in Information Technology, is available now. He may be reached at cpickering@sdireports.com.

Editor's note: This column first appeared on Datamation, an internet.com site.