IBM’s generative AI tool aims to refactor ancient COBOL code for its mainframes

Photo of author
Written By Editor

Lorem ipsum dolor sit amet consectetur pulvinar ligula augue quis venenatis. 

COBOL 73 running on an IRS computer in an emulator

There are hundreds of billions of lines of COBOL code operating on production systems worldwide. That’s not perfect for a language over 60 years old and whose main designers are mostly retired or dead. IBM, excited to keep those legacy works on its Z mainframe systems, wants that code rewritten in Java. It attempted getting humans to do it a few years back, but now it has another

concept. Yes, you guessed it: It’s putting AI on the task. The IBM watsonx Code Assistant, slated to be readily available in Q4 this year, plans to keep human beings in the mix, but with a push from generative AI in analyzing, refactoring, and checking the new object-oriented code. It’s not an all-or-nothing process, either, as IBM claims that watsonx-generated code should be interoperable with COBOL and certain Z mainframe functions. In a technical article particular to COBOL conversion, IBM’s Kyle Charlet, CTO for zSystems software application, steps up to the plate and says what a great deal of individuals have actually said about COBOL: It’s not simply the code; it’s business reasoning, the edge-cases, and the institutional memory, or the lack thereof. Some clients … have actually spent years and millions attempting to update their COBOL code, and despite all that, just a fraction has actually been updated. It takes great deals of developers and painstaking manual work to reword COBOL. People have attempted various ways to update these applications, and in my experience, with diverse results. None are exemplary. IBM’s watsonx, Charlet composes, might assist large organizations decouple individual services from monolithic COBOL apps. In IBM’s view, it will come in 3 steps: Refactor, in which specific services are”surgically”separated or extracted from bigger code Change, either to mainframe-friendly Java code or COBOL that can speak directly to Java Validate, with AI helping to produce test cases but with coders still”in the motorist’s seat ” In some ways, AI support looks like it can just help with the generational obstacle of moving COBOL forward while keeping it practical. While COBOL codebases

  • can be safe and relatively stable– as soon as discovered to be among the least problematic in a broad survey– the costs of
  • updating and extending them are enormous. Tradition COBOL was among the factors the Office of Personnel Management suffered a deeply invasive break-in in 2015, as the old-fashioned code might not be encrypted or made to

deal with other secure systems. But there’s a repeating argument that COBOL is good at handling business-specific systems and exchanges in ways that(some may argue)present less attack vectors. Or you may argue that AI-generated and restructured code might look appropriate and seem test-ready, but without individuals around who understand exactly why the code does the things it does, AI-upscaled code could have just as much sound as AI-upscaled video. IBM’s watsonx Code Assistant for Z will next be deployed for Red Hat Ansible Light speed. Given that was purportedly trained on more than 100 coding languages, more AI co-pilots for old mainframe code will likely follow.

Leave a Comment