Before you rewrite the application, decide what deserves to survive.

Automated rewriting has changed what is technically possible. Large applications can be translated, tested and moved with far less manual effort than earlier generations of transformation programs required. That is meaningful progress. It does not remove the need for a business decision.

Behavioral equivalence answers a technical question: can the new implementation produce the same result as the old one? It does not answer whether the behavior is still wanted, whether the policy remains valid or whether an exception should be carried forward.

Preserving behavior is not the same as deciding that the behavior deserves to survive.

The application is carrying business history.

Long-lived systems accumulate more than code. They contain pricing rules, eligibility decisions, regulatory obligations, operational workarounds and terminology that may no longer appear in current documentation. Some are essential. Some are obsolete. Some conflict with how people now believe the business operates.

A rewrite factory is designed to preserve equivalence. The organization must decide what to keep, change or retire. That requires a view of the application that business and technology participants can examine together.

Separate business intent from implementation.

Traditional code analysis is useful for understanding structure: components, calls, dependencies, data movement and technical risk. Discovery 1.0 addresses a different problem. It recovers business rules, terminology, processes and data relationships, then keeps those findings connected to source evidence.

This separation gives a team something more useful than a logic dump. It creates an evidence-backed business model that people can review before a target platform or delivery method makes the decisions expensive to revisit.

Begin with one consequential decision.

The first step does not need to be an enterprise-wide transformation. Begin with an application and a question that matters now: a looming retirement, an unexplained behavior, an audit request, a consolidation decision or a planned rewrite.

The goal of the first review is not to prove a platform can analyze code. It is to discover whether the recovered knowledge changes the decision in front of the organization.