Understanding and Integrating
Three experts discuss the value of legacy applications and how they can move business forward
Illustration by Russell Charpentier
Legacy applications. To some, the phrase brings to mind old, out-of-date software written in decades past in dead programming languages by people who no longer exist—and to some degree, they might be right. Many of these applications were in fact developed years ago and the programmers who created them may be long gone, having either moved on to other jobs or retired.
But that makes this software no less important to the organizations that have come to rely on it. Often, “legacy” describes the battle-proven applications that have kept things humming from their initial launches, through multiple alterations, to today. They’re mission critical and up to the challenge.
Unfortunately, the people who understand these applications are slowly and quietly fading away, leaving behind millions of lines of code that younger, graphically driven programmers may not entirely understand.
To counter this, many vendors provide tools that let these younger programmers delve into the inner workings of legacy applications and then make them applicable to the modern computing environment.
To learn more about how IBM* System i* users can ensure that their legacy applications can meet the challenges of today’s—and tomorrow’s—computing requirements, we spoke with several experts in the field, including Bob Cancilla, System i software evangelist, IBM Rational* Tools; Joel Eikenhorst, president and CEO of consulting firm iCAN; and Stuart Milligan, vice president of business development with Databorough Ltd.
IBM Systems Magazine: What exactly are legacy applications?
Milligan: The term “legacy” can be a misnomer, and the problem is that the word is sometimes used to describe something that’s old and therefore decrepit. But it can also mean a massive, rich inheritance. So when you talk about legacy applications, it’s like an inheritance that drives your business.
Eikenhorst: Without them, businesses just couldn’t continue to run. And they’re not just back-end accounting apps. They’re usually either tracking all of the sales or they’re tracking all of the shipments and things like that. So in my mind, “legacy” means applications that are critical to the business.
Cancilla: Let me add to that by saying that they basically fall into two categories: first, ISV-provided software purchased from a third party and customized and extended by customers, or, second, those that are developed in-house. One of the issues that we see in terms of legacy apps is they’re often very old. Some of them date back 40 or 50 years when the initial development was done. And the real issue with them is the fact that they’ve been modified constantly over time to address big issues like Y2K and, more likely, the day-to-day changes in business operations. And in a 20-, 40- or 50-year timeframe, there’ve been a lot of developers working on them.
Eikenhorst: That’s true. In our experience, most applications keep evolving. Developers keep adding on and adding on, doing some modernization on pieces of it as they need to, or helping integrate them with other applications.
IBM Systems Magazine: So it sounds like over the years, modifications might be one of the biggest issues regarding legacy applications.
Eikenhorst: That’s true to a point, but there’s more to it than that. CIOs realize that they’ve got a problem here, and that they can’t simply go out and find the people with the business knowledge that their existing developers have. That’s something they can’t replace. So it’s not just the application modifications, but also finding people who understand the business and how all of these applications are put together to support the business.
Milligan: Within almost every System i installation, you can find someone who can write RPG and knows quite a lot about the business. That’s in part because the RPG code describes the implementation of business processes, and the people who manage the applications actually understand the business—more so than in, say, Microsoft* technology-centric companies. When you go into a Visual Basic shop and ask, “Why does this warehousing and logistic application have a 15-digit code?” the likely response is “I don’t know.” You ask RPG programmers, and they’ll give you a description of the history of the business.
Cancilla: Unfortunately, we’re seeing a rapid decline in this RPG programming population. I talk to CIOs on a daily basis from customers all over the world, and everyone is concerned about staffing and the ability to hire RPG programmers. They’re extremely concerned about current programmers retiring within a five- to seven-year period.
Eikenhorst: It’s important to understand, though, that RPG has evolved quite a bit so that it’s now a modern language. You can do about anything you can do with most other modern languages, but a lot of people just haven’t taken advantage of what’s there because they’ve been concentrating on running the business. But they also have to acknowledge that they have to integrate their legacy applications with external businesses, in an SOA (service-oriented architecture) or Web-services environment.
Cancilla: That’s right. For example, if you’re a manufacturing company that builds something and sells it to a big-box retailer, you’re going to be involved in SOA and Web services whether you like it or not. If you want to do business with them, you’re going to implement systems that meet their requirements. They’re saying, “This is how we do business and we have to have electronic interactions with you—we’re going to communicate over the Internet. If you don’t want to play, we’re not going to play with you.”
Milligan: Going back to an earlier point, it should be noted that RPG isn’t dead. The fact that it runs 85 percent of every financial transaction that takes place in the world every day means that the predictions of RPG’s demise are glib statements that shouldn’t be taken too seriously.
“RPG has evolved quite a bit so that it’s now a modern language. You can do about anything you can do with most other modern languages, but a lot of people just haven’t taken advantage of what’s there.”—Joel Eikenhorst, president and CEO, iCAN