At Mechanical Orchard, we’re focused on tackling the most critical and complex legacy modernization efforts.

‍

The practical demands of safely rewriting and moving a legacy application call for a different approach from traditional code transpilation or “big-bang” re-architectures of business processes. Rather, we start with the behavior of the working system as it currently is.

‍

This is a safer approach, one that is designed to increase confidence and reduce risk through demonstrating meaningful progress at every step. We follow three principles:

‍

• The running system is the specification
• Reverse-engineer and create a behavioral twin
• Deploy incrementally into production

‍

Here’s how it works.

‍

Determining where to start

Legacy systems often present as vast monoliths comprising tens of millions of lines of code, which begs the question: where to start? What “seams” can we identify to discern relatively discrete components? What other system elements do these components interact with? The more critical the system, the more important comprehension is.

‍

‍

Tackling the first workload

Once we’ve selected a workload, we analyze its data flows, documentation, user interactions, and code. We synthesize this information to form a behavioral model, creating a test suite to accurately capture the workload’s behavior.

‍

‍

Delivering the first workload

We build confidence—and increase control over the success of the modernization effort—by focusing on proving equivalence and getting incremental workloads into production quickly.

‍

‍