The interesting part
Every new project brings a temptation to try the framework everyone is talking about. We almost always resist it.
The products we build will be maintained for years, mostly by people who were not in the room when the choices were made. The best gift we can leave them is a stack they already know.
What we choose
- TypeScript on the web, with a framework that has been stable for years.
- Swift and Kotlin on phones, written the way their platforms recommend.
- Postgres behind them, because it does nearly everything and everyone can hire for it.
The interesting part of a product should be what it does, not what it runs on.
When we make an exception
When a product genuinely needs something unusual, such as real-time location for Ferrow's drivers, we use the boring tool for everything else and isolate the unusual part behind a small, documented boundary.
