Here are some points to ponder. First, the RPG programmer pool appears to be shrinking while the average RPG programmer age is rising. Second, the AS/400 is losing market share in some key areas like Enterprise Resource Planning (ERP). For example, in the June 2000 MC article “ERP Sales Rebound, But Not as Expected Thanks to E-biz,” Timothy Prickett Morgan reported a nine percent loss in the AS/400’s share of the ERP market from 1996 to 1999. Third, shops are moving to a mixed environment where the AS/400 is just one of the platforms in the corporate network.

These points and others tell me the days of green-screen RPG programming are numbered. While I’m not suggesting you panic, there is no denying that the corporate environment is evolving, and the AS/400 is evolving right along with it. The trick is for IBM to pace the evolution so that it remains a meaningful force in the industry but does not leave behind its customers. The trick for managers is to be able to cut through the hype and know when to fold these new technologies into their corporate infrastructure.

One promising new technology is the Java/servlets/JavaServer Pages (JSP) architecture. Should a shop with some sharp RPG programmers start using Java/ servlets/JSP for all new development?

Let’s consider the impact of such a change. Java/servlets/JSP performance is still slower than the green-screen RPG world. This shop will either need spare processing capacity on its existing AS/400s or have to consider getting a larger box. Make a $200,000 investment in processor upgrades if the result will be an application that performs about the same as the current application, only now with a GUI or Web interface?

The next impact is infrastructure and setup. WebSphere takes some time to set up. Getting everything configured just the way you want it takes several months if you have to learn as you go. The shop is going to have to dedicate one or two people for several months to get everything ready for the other developers.

The entire organization must shift its mindset from RPG to Java. The sharp RPG programmers will have to learn both Java syntax (easy) and object-oriented concepts (somewhat harder). Meanwhile, the rest of the company will have to understand that the programmers don’t have time to create that neat new report for the payroll department.

Another impact is the requirement of client workstations. Is your company ready to get rid of all 5250 dumb terminals? If your company has a dirty warehouse, will it get dust- tolerant thin-client workstations or PCs that won’t crash after a couple of months?

Finally, consider the change in programmer productivity. Programmer productivity is going to be a great deal slower for some time after moving to Java servlets. You are looking at doubling the time required to do the equivalent in RPG. Management is going to have to get used to that one. The benefits of objects and reusable code don’t show up until you have built a critical mass of classes and the programmers have gotten some experience under their belts.

If all this has you thinking twice about moving to Java, consider the pool of new programmers. Compare the number of available Java programmers to the number of RPG programmers, and you will find the reason this transition is happening, regardless of the issues mentioned above: supply. In the short term, I see companies spending large sums of money getting software vendors to make RPG changes on a time/materials basis. This will be true even for on tasks that used to go to very junior staff programmers.

This kind of anecdotal evidence convinces me that the RPG programmer shortage will continue to get worse. Companies may be willing to work around it in the short term, but for the long term, RPG is a dead end. I really don’t worry about the companies. They will eventually make the hard decisions that keep them in business.

I worry about the programmers who have spent years building RPG skills. It is already a common business practice to dump aging (i.e., expensive) programmers in favor of younger (i.e., cheaper) programmers. If the programming skills of the older programmers are also obsolete, well, that can’t help.