The Good News is that a very large number of the COBOL programs work as well as they ever going to. And they would work just as well if Recompiled for more flexible cost-effective systems. Yes, the COBOL is only a piece of the puzzle. Figuring out what to do with related materials (JCL, utilities, third-party packages) may be more difficult than the recompile.
The easy programs are the background or “Batch” programs that are run by automatic scheduling programs and are very “hands-off”. The programs that run on mainframe Online Transaction Processing systems (like CICS) could be recompiled to run on Open Systems workalikes like OpenKicks. Or they could use APIs from Web or Streaming front-ends. COBOL supports XML and JSON, two popular API data-protocols.
Recompiling needs just as much testing as a rewrite, of course, so a project is not a quick, cheap, and risk-free undertaking. And you end up with COBOL that is just as easy to maintain and extend as the original. That’s typically not true for machine programming language to program language translation. And there is no reason to believe AI is going to do any better. Maybe it will but we would be surprised.
With the COBOL programs running on whatever modern platform (in-house servers, VPS servers, containers in The Cloud), the longer strategic projects can avoid “Big Bang” all-or-nothing approaches. Those high-pressure methodologies lead to too many outright failures and/or large cost overruns.
About the COBOL Skills Shortage
This is a problem. It has a problem we have seen coming for years. Many enterprises have recognized this challenge and pro-actively recruited younger developers, trained them, and will have no such problem. Others have not. Either in part or in whole. And all too often they are reminded of the systemic importance of the COBOL programs when a key older COBOL developer retires.
Marist College, North of Poughkeepsie, NY, and Northern Illinois University are two schools with current mainframe-oriented programs and COBOL courses. Marist’s program was well funded by IBM executives who lived and worked nearby years ago. NIU was a surprise at a recent users group meeting. We all cheered. Programs like that have disappeared because educational institutions didn’t see demand for Business IT skills.
Developing and teaching COBOL courses is not so hard. There are companies who sell the service still. The issue is that firms facing COBOL skills shortages need to make the need for training visible in the marketplace so supply will improve.
