Built by people who have been boarded.

Why most of it does not fit a ship

Most software written for ships is written for the office that buys it. You can tell within about thirty seconds: it assumes a connection, it assumes a keyboard, it assumes somebody has time to fill in fields that exist to make a report render, and it assumes the person using it cares about the report. On a vessel, none of those are true.

What that changes

Offline is the normal state, not the error state. Every screen a crew member touches works with no signal, and nothing is lost waiting for one. The synchronisation is designed around the assumption that the link will fail in the middle of a transfer, because it will.

There is one record, not a vessel copy and a shore copy. The finding the crew raised and the finding the superintendent opens are the same object. Everything on this website that claims two things are connected means connected in that sense — and the site carries an automated check that fails if any page starts describing a record instead of reading it.

And the product does not flatter the fleet. It will show a vessel nobody has heard from as unknown, it will not invent a severity score, and it will not let an inspection item close because somebody ticked it. Software that makes a fleet look better than it is has an obvious commercial appeal and is, in this industry, a safety hazard.

A system that is comfortable to use and wrong is worse than no system. The crew already know what is wrong with their ship.