Small Systems

I do not want a system for everything. I want the smallest system that makes repeated friction disappear.

This page collects small systems I build or test for repetitive work, context switching, clutter and problems that are annoying enough to deserve a fix but not big enough to deserve a giant project.

Start with repeated friction

I usually get a better result when I stop asking “What should I automate?” and start asking “What keeps making me do the same avoidable work?”

Field note: I Keep Building Tiny Systems for Problems Other People Ignore

Compress the problem before building the system

A useful small system needs a narrow target. If I cannot explain the problem clearly, I am likely to automate the wrong thing or create a new layer of maintenance.

Field note: If I Can’t Explain the Problem in One Sentence, I’m Not Ready to Build

Do not automate a feature. Remove a friction.

A feature list can grow forever. A named friction gives me a way to reject ideas that do not make the original problem smaller.

Field note: Most Product Ideas Start in the Wrong Place

Keep systems small enough to delete

A tiny system should earn its maintenance cost. If the workaround becomes more complicated than the friction it removes, simplification has turned into another project.

The decision rule I keep using

Automate repeated, predictable friction when the result is easy to verify. Leave the rest manual until the pattern becomes clearer.