IBM is developing a generative AI-assisted software development tool, watsonx Code Assistant for Z, to help developers translate Cobol to Java on IBM Z. The tool uses IBM’s watsonx.ai code model to modernize legacy Cobol applications. The tool promises to produce object-oriented Java code optimized to interoperate with the rest of the COBOL app with the Customer Information Control System (CICS) transaction processing subsystem, Information Management System (IMS) transaction processing database, Db2 transaction and analytics database, plus other z/OS operating system runtimes, according to IBM.
IBM’s Institute for Business Value found that nearly 70% of IT executives consider mainframe-based applications central to their business and technology strategies. However, almost 70% of executives surveyed reported that the mainframe-based applications in their organizations need to be modernized.
The tool will be previewed during IBM’s TechXchange event in September, with general availability expected in Q4 2023. IBM expects partners of all types, including system integrators, value-added resellers, independent software vendors, and hyper scalers, to play a role in bringing the solution to market.
The model was trained on IBM’s CodeNet, a database of 14 million code samples in over 55 programming languages, and was tuned for two specific use cases: a coding assistant for RedHat’s Ansible automation toolkit and the new COBOL solution.
I’m focusing on practical applications in business for AI. Code modernization seems incredibly valuable. The implications for catching up on technical debt here seem critically important. Translating an aging codebase to a modern platform is a slam dunk. This application to old code bases that teams become wary of touching – is addressed by the application of AI, which can explain and trace code. The opportunities here are significant.. which is why we care.