Hi everyone,
Thanks to everyone who joined our recent webinar about ACUCOBOL x GenAI (watch the recording) and shared such thoughtful questions. We received a wide range of topics from UI controls to AI‑assisted development workflows, modernization and code analysis.
Here are the webinar questions and answers from our team. If you have any other questions, let us know in the comments.
Question: I have seen elements for maximize and minimize in the screens—are these fully functional in ACUCOBOL?
Answer:
Yes, they are fully functional.
When using Rocket AcuToWeb, the Web.Resize configuration in the gateway.toml file can automatically maximize your screen based on the ratio you define. This allows the application to adapt cleanly to different screen sizes.
At runtime, screen elements can be repositioned dynamically, although this does require some additional coding. If you prefer to avoid manual repositioning, version 11 introduces NTF‑SCALING, which automatically resizes screen elements without extra development effort.
Question: How can we embed AI‑assisted understanding, navigation, and explanation of large ACUCOBOL codebases directly inside the VS Code developer workflow—without requiring developers to leave the IDE or rewrite COBOL?
Answer:
There are a few effective ways to do this today. If you are using GitHub Copilot in VS Code, you can select any portion of ACUCOBOL code, right‑click, and choose “Explain” from the contextual menu. Copilot will generate an explanation directly in the Copilot chat panel within VS Code.
You can then copy that explanation into your source file as comments, meaning your code becomes documented without ever leaving the IDE. Additionally, the Rocket ACUCOBOL‑GT Extension for VS Code allows you to quickly comment or uncomment blocks of code using Shift + Alt + A, making it easy to integrate AI‑generated explanations into your programs.
Question: Can Rocket ACUCOBOL’s Gen‑AI framework be used as a supported modernization platform to migrate Mainframe workloads—such as Assembler programs or JCL/Quick Jobs—into ACUCOBOL (or COBOL on distributed platforms)? Or is it limited to AI‑assisted understanding of existing applications?
Answer:
If your goal is to migrate Mainframe workloads such as Assembler programs or JCL, we recommend looking into the Rocket Enterprise product line, which is specifically designed for that type of modernization.
Gen‑AI capabilities shown on the webinar are primarily focused on understanding, modernizing, and improving existing ACUCOBOL and COBOL applications, rather than serving as a full Mainframe migration platform.
Question: Companies are looking at AI as a way to mitigate COBOL risks—especially knowledge retention and long‑term maintenance. Do these features make a strong case for keeping COBOL functionality into the future?
Answer:
Yes. We see AI as a powerful productivity and knowledge‑enablement tool.
When combined with tools like Rocket COBOL Analyzer, AI helps developers better understand code, reduce onboarding time, and maintain complex applications with more confidence. Rather than replacing COBOL, AI strengthens its long‑term viability by making systems easier to understand, maintain, and evolve.
Q: Will COBOL Analyzer identify fall‑through paragraphs correctly and not mark them as dead code? Example:
A section is performed, then execution flows through several paragraphs until an EXIT.
Or a PERFORM A01‑START THRU A99‑END statement, where multiple paragraphs in between are not directly called.
Answer:
Yes, Rocket COBOL Analyzer correctly identifies fall‑through paragraphs.
These paragraphs are treated as valid, not dead code. When visualized, they appear connected by a red line, indicating fall‑through logic. For reference:
Blue lines represent external calls
Brown lines represent PERFORM statements
Green lines represent GOTOs
A paragraph is only marked as dead code if it is truly unreachable, for example:
A preceding paragraph contains an EXIT
The only PERFORM calling it was commented out
The calling logic was removed entirely but the paragraph was left behind
Question: How well can COBOL Analyzer detect inline code that is skipped and will never be executed?
Answer:
Rocket COBOL Analyzer builds a control‑flow graph (CFG) and performs thorough static analysis.
If a block of code is impossible to reach, it will be identified as such. Detection is especially strong when conditions are based on constants. However, if execution depends on external or dynamic values, COBOL Analyzer will treat the code as potentially reachable, even if it may not be in real‑world scenarios. This ensures accuracy without making unsafe assumptions.
Question: Where do APIs or MCP fit into the product roadmap?
Answer:
We are actively integrating AI across our product portfolio.
In Rocket COBOL Analyzer, we are adding LLM support in the Chat section, allowing users to request insights and information directly within the tool. We are also configuring an LLM to generate CSS for Rocket AcuToWeb via chat, providing an alternative to the traditional Theme Generator.
MCPs are primarily aimed at AI creators and framework developers. Our focus is on configuring and integrating AI technologies so they deliver practical, high‑value assistance to our users without requiring you to manage AI infrastructure yourself.
