Legacy code is the old software that still runs important business operations but has become hard, slow and risky to change. It is not always about age: a system becomes "legacy" when no one understands parts of it anymore, when it lacks tests that make changes safe, or when the technology lost support.
Why ignoring it is dangerous
Legacy code often holds business rules that exist only in that code, with no documentation. Every change becomes a risk, new developers take forever to be productive, and the company becomes hostage to a few people. Ignoring the problem does not freeze it: it gets worse, because technical debt keeps accumulating.
What to do
Before choosing between refactoring, rewriting in parts or rebuilding, the right step is to diagnose the code's real state: debt, test coverage, obsolete dependencies and risk. Only with that map does the decision stop being instinct and become strategy.