Understanding technical debt and legacy code

While computer science education, books, tutorials, and online courses all focus on creating new projects from scratch, the reality is that almost all development jobs you’ll have will center around understanding, maintaining, and expanding pre-existing code that may not meet your current standards.

This pre-existing code is referred to as legacy code. You almost always inherit some amount of legacy code when joining a new project. This can be a large amount of code for pre-existing projects or a smaller set of libraries your code must work with.

There are many different definitions of the term legacy code. One that stands out to me from my readings is Michael C. Feather’s definition, in Working Effectively with Legacy Code, that legacy code is code without tests.

While I like Michael’s definition and believe testing is critically important, as we’ll see in Part 2 of this book, I personally define legacy...