Abhijit Deshpande Notes
← All notes

Engineering ·

The case for boring, reliable technology

A personal project should spend its complexity budget on the problem it solves, not on the infrastructure that runs it.

Both of my live websites run on almost nothing. This one is plain HTML, CSS and a few lines of JavaScript on GitHub Pages: no build step, no framework, no database. nihiramusic.in does a little more — its content lives in JSON files and a handful of Python scripts run on a schedule — but there is still no server to patch and no hosting bill.

What boring buys

Nothing to keep alive. Static files don’t crash at night, need security updates or run out of memory. The hosting is someone else’s well-run problem.

Every change is visible. Content and code live in a Git repository. Each change is a readable diff, and if something goes wrong, the previous version is one revert away.

It still makes sense later. In six months I will have forgotten the details. A plain HTML page explains itself; a stack of tools and configuration needs to be relearned before it can be changed.

It costs nothing to run. That matters for side projects, which tend to be abandoned the moment they become a monthly bill.

When to add moving parts

Boring doesn’t mean never adding anything. nihiramusic.in had a real requirement: it should keep itself up to date when new videos and concerts appear, without anyone remembering to edit it. The answer was not a backend, but scheduled jobs that update data files in the repository and redeploy the same static site. The moving parts run on a timetable and leave a trace in the history.

The rule I try to follow is simple. Start with the simplest system that meets the requirement. Add a moving part only when a real constraint asks for it — and prefer the kind you can see, review and switch off.