Hi everyone,

The idea that COBOL is "stuck in the past" is one of tech's most persistent myths - and one of its biggest disconnects, given how much COBOL is still actively being developed.

Even those once familiar with the language can struggle to see how it fits into a modern tech stack in the age of AI. Perhaps you're one of them?

There's more than a grain of truth, of course - many COBOL systems do carry the weight of their history. Real-world COBOL often contains layers from different eras, as you'd expect from software that's delivered value for decades.

But to examine a few lines of code written from 1980s or earlier and conclude COBOL has no place in modern IT is like dismissing smartphones because you once used a brick-sized mobile phone in the '80s.

Don't let your memory of COBOL - as it once was and when you once wrote it - cloud your judgment of what it's capable of today. And perhaps most regrettably, don't let that lead you to trade it for something that's little more than the emperor's new clothes.

COBOL has evolved. Take this snippet: a Fibonacci demo in COBOL using local variables, parameterized sections, and recursion:

 *> COBOL 

       program-id. Fibonacci.

       procedure division.
           display fib(10)
           goback.

       fib section (n as binary-long) returning result as binary-long.
           if n <= 1
               move n to result
               exit section
           end-if
           compute result = fib(n - 1) + fib(n - 2).
       end program.

But does any of this actually matter?

Well, no one's suggesting we start building brand-new global banking systems in COBOL. But the ones that already exist - and as you know, there are a great many of them - they need to be understood, maintained, and extended.

That's why COBOL language evolution matters - for business continuity and for bringing in new talent. Its evolution is ongoing, year after year, without forcing you to rewrite the past to embrace the future. If you're coming from Java, Python, or C#, you'll recognise a lot here. The learning curve is more of a gentle ramp than a cliff.

In an age of AI, there's something remarkable about a language like COBOL still evolving. At the ripe age of 63, this dog is still learning new tricks.

And as software development becomes more agentic in nature - with AI and automation taking a more active role in coding - these modern constructs become more than just conveniences. They enable new kinds of tooling, code generation, and automated maintenance. A more expressive COBOL becomes a more usable COBOL - for people and machines.