Why PL/I deserves its own modernization plan
PL/I is one of the languages Replai reads. The platform builds a structured map of what a system actually does, covering business rules, workflows and dependencies, across COBOL, Oracle Forms, PL/I and others. The founding team spent years inside enterprises running on systems that no single person fully understood. Core banking platforms, insurance policy engines and logistics infrastructure came up again and again, built from decades of business logic written in COBOL, PL/I and aging Java codebases, with the original authors long gone and the documentation missing or wrong.
This guide is for teams that own PL/I code and are weighing a PL/I modernization. It covers what to settle before you pick a target, and where discovery-first analysis lowers the risk.
Where PL/I modernization projects get stuck
The pattern is the same one we see across legacy languages. Teams start with good intentions, then spend months just trying to understand what the existing system does before any new code is written. Some projects fail outright. Others deliver a fraction of what was promised, because critical business rules stayed buried and got lost in translation.
For a PL/I migration, that tends to show up in a few ways:
- Nobody holds the full picture. The people who wrote the code have left, and what documentation exists may be incomplete or wrong. See what to do when the only person who understands your system is about to retire.
- Business rules are implicit. A rule that lives in code and nowhere else is easy to miss and easy to lose in a rewrite.
- Dependencies are unclear. Changing one field can affect components you did not know were connected. A change impact analysis done before the work starts is cheaper than one done after something breaks.
- Line-by-line translation carries the past forward. Converting syntax alone moves old structure into a new language without recovering the reason the code exists.
Start with discovery, not translation
Replai takes the view that a line-by-line rewrite carries the past forward, while a semantic knowledge layer gives you the understanding to choose a better future. For a PL/I codebase, that means reading the source first and recovering the meaning before you decide how to rebuild.
In practice, discovery produces four things your team can use:
- A map of the business rules, workflows and relationships in the code, each linked to its source evidence.
- Plain-language answers to questions about what a program does, so people who did not write it can still work on it.
- A view of how a change to one field ripples through the system. Our guide to data lineage in legacy systems shows how to trace a field from screen to table.
- A foundation for planning the migration with business intent, not just syntax, as the starting point.
Human judgment stays in the loop throughout. The analysis gives your architects and developers evidence to decide with. It does not replace their decisions. For a plain-language look at the mechanics, read how AI agents read legacy code.
Choosing a target such as Java
Many teams searching for PL/I to Java are really asking two questions at once: what does the system do today, and what should it become? Answer the first before the second. Once the rules and dependencies are visible, you keep your options open and can rebuild in the architecture and technology that fit your future, rather than the one a translation tool happens to produce.
If your estate also includes COBOL, the same discovery-first approach applies. The COBOL to Java migration guide walks through discovery, business logic preservation and testing, and much of that planning carries over. Where a project runs on IBM mainframes and is heading to the cloud, see mainframe to cloud migration. Teams unsure whether to modernize or replace can compare the two in AI-driven modernization vs. full system replacement.
What to do before you start
Before you commit to a PL/I migration plan, get clear on a short list of things:
- Which programs and data stores are in scope, and who still understands them.
- Which business rules must survive the move unchanged.
- How you will verify behavior after the move, so you are not relying on memory.
- What your budget assumes about discovery. Missing documentation is one of the biggest cost drivers, covered in legacy system modernization cost.
If you want to see how this works on your own code, you can explore the Replai approach to discovery and modernization and talk to the team.
PL/I modernization FAQ
Why do PL/I modernization projects stall?
Teams often spend months just working out what the existing system does before writing new code. The original authors are usually gone and the documentation is incomplete or wrong, so critical business rules can stay buried and get lost in translation.
Should I translate PL/I to Java line by line?
A line-by-line rewrite carries the past forward without recovering why the code exists. Recovering the business rules and relationships first lets you rebuild in the architecture that fits your future.
What does discovery-first analysis involve?
It reads the existing code and builds a structured map of the business rules, workflows and dependencies between components. Your team can then ask plain-language questions about the code and trace how a change will ripple through the system.
Does Replai support PL/I?
Yes. Replai reads existing code in COBOL, Oracle Forms, PL/I and other languages to map what the system does.
Does AI replace the engineers on a PL/I migration?
No. Human judgment stays in the loop. The analysis gives architects and developers evidence to make decisions, rather than making them.
Is the approach for PL/I different from COBOL?
The discovery-first approach is the same across languages: understand the business meaning in the code before you change it. Guidance on COBOL to Java migration carries over to much of the planning.