Thread

Okay, let's talk about "duplicate work." I've often heard the argument "let's not do duplicate work" to keep teams from doing something. And it seems to make sense. Why re-invent the wheel? But in reality, the picture is often more complicated.

First, what is considered duplicate work often turns out to be not exactly duplicate work on closer inspection. So superficially it looks like the same thing, but in reality these are slightly different and incompatible things.

Often, the work supposedly being duplicated hasn't happened yet. In that case, you couple two streams of work that were unrelated before. This leads to coordination and synchronization overhead, meaning it will take longer.

Sometimes, the de-duplication requires to build a slightly (or majorly) more generalized piece of software. Which might not make sense or be feasible. It is also not "the simplest thing that works."

Lastly, telling a team not to do something they absolutely feel like they need and want to do can be quite motivating.