Perfection blocks progress
A project like Ivrea Observatory can easily become endless: choosing the best sensor, rebuilding the UI, changing architecture, waiting for the right camera.
The healthiest choice is to make it work early, then improve it. An imperfect dashboard fed by real data is worth more than a perfect project never published.
Every version teaches something
The first version teaches whether data arrives. The second whether the page is readable. The third whether the sensors are stable. The fourth whether the history tells something.
Even errors are part of the method: hosting cache, UTF-8 encoding, wrong paths, sensor false positives. Everything becomes material to make the project more robust.
The journal is useful for this too
Writing the journal forces decisions to be clarified. Why is the dashboard the homepage? Why is lightning not in the menu? Why is the 21 June claim in the footer?
The answers make the project more coherent, not just prettier.
Highlights
On the workbench
- Keep the project moving without losing the reasoning behind each step