Don’t build systems, build subsystems
blog.jessitron.com
blog.jessitron.com
There is an anti-pattern, counter to this advice, which I call software-psychopathy.
This is when someone (the psychopath-developer) surveys the entire problem landscape of the full system and identifies a sub-problem, carves out a sub-system that is unnatural but fits their competency and through their position in the software-social-order and/or through personal charisma pushes a solution which makes all surrounding problem space even harder to solve.
The psychopath-developer says, "I see we have problems, A, B, C and D. I will solve B with X and now all that the solutions to A, C and D need to do is satisfy requirements L, M, N, O and certainly Q. Aren't I a genius? Now, excuse me as I leave the details to you and go "help" another group!"
A more charitable term might be "picking low-hanging fruit". Inside this metaphor, by thoughtlessly picking the heavy low-hanging fruit the psychopath-developer unweights the branch sending the rest of the fruit higher, out of easy reach and adding burden to those suckers that come later to actually get all the fruit. Meanwhile, the psychopath-fruit-picker is long gone enjoying the fruits of his easy labor.
1. https://en.m.wikipedia.org/wiki/Rule_of_three_(computer_prog...
A junior will create the thing or carry out the task that's asked of them with but are often tone deaf to the context that surrounds the request and outcomes desired. A senior will implicitly addresses the wider context, asking probing questions to that end, without prompt.
You can also apply this concept in the opposite direction, to aid in understanding systems you encounter. Considering it within the context it operates in, but also what the intentions were of whomever created/deployed it. Considering both angles usually leads to far better understanding in less time.
While this doesn't market well, it is a better tool for thinking. Not creating artificial intelligence, but contributing to smart systems.
Write programs that do one thing and do it well.
Write programs to work together.
Build for platforms, don't build in a vacuum. Ie. Use an existing battle-tested framework if you can.
And.. Use existing subsystems if they already exist instead of reinventing the wheel.