[bgribble]

There's no hurt feelings like the hurt feelings of a junior engineer, who has spent the last year kvetching about how much they hate working on legacy junk, hearing someone else refer to one of THEIR projects as "legacy junk".

Any code that's old enough to have its first birthday party is "legacy", which means that "legacy" is a completely useless category. Anyone calling anything "legacy" is generally just showing their own lack of experience.

[milesrout]

1 year old code is not legacy. Legacy is a useful category. COBOL is legacy. Maybe it is unclear exactly where to draw the line but if that were a valid reason to discard conceptual categories we wouldn't have any.

[jaggederest]

I like the definition from "Working Effectively with Legacy Code", Michael Feathers says legacy code is code without tests.

[sethammons]

Heh, so true in many cases. I had rewritten a utility that scanned a directory and moved files to s3 from Perl to Go and eventually a different team took over the code. A bug popped up and they were not confident to update the "legacy code." I could do not but chuckle to myself: this was like a fee hundred lines of well organized and fully unit and integration testable code with a readme and runbook and grafana/prometheus and structured logs aggregated in Splunk. And that has been running for like 2 years. They just didn't want to even attempt the fix and in fact pushed it off indefinitely.

[protonbob]

Legacy code is code without tests. Doesn't matter if it's new or old.

[Boldened15]

IMO legacy code is code that has lots of if statements and special cases sprinkled in over the years to enable new features quickly and get out immediate hotfixes for bugs without doing full refactors and/or data migrations to address the root causes. Even with 100% test coverage it’s a pain to build new features in because there are so many paths to think through, and instead of single sources of truth each part of the app assumes every other part is working in very specific ways.

An alternative definition, legacy code is any code where there’s no one left on the team who has been working on it for years and intuitively knows the pitfalls. Then everyone’s scared to touch it or make big refactors, which actually leads to those small special cases being added instead.