What I wish I knew before going fullstack
Placeholder article — every post currently renders this same body. Swap it for real content per slug before publishing.
The first version of this shipped on a Friday afternoon and broke on Monday morning, which is the part nobody writes down. What follows is what I would tell myself before starting, in the order the problems actually arrived.
Start from what breaks
The interesting constraint was never the framework. It was that two systems had to agree on the same numbers while neither could stop to wait for the other, and every design that ignored that fell over the moment real traffic arrived.
Every abstraction I added to feel safe turned out to be another place the truth could go stale.
So the second attempt kept one source of truth and made everything else derive from it. Less clever, much shorter, and it survived the week.
What I would keep
- Write the failure case down before writing the happy path.
- Measure before optimising anything, including the thing you are certain about.
- Leave the seams visible — future you needs somewhere to cut.
None of this is novel. It is just the shortest path I found through a problem that looked much bigger from the outside, and it is the version I wish I had read first.