I have direct experience of that specific implementation and you are correct in that it was exceptionally mis-managed. Even in its inception, it was stupid, so

I have direct experience of that specific implementation and you are correct in that it was exceptionally mis-managed. Even in its inception, it was stupid, some company sells to Melbourne and outsources the business logic to Switzerland and outsources the rest of the app to a UK firm. Idk if you've ever worked on embedded but having to deal with an opaque third party component that is doing all the "business logic" and hogging resources (SqlCe makes this _really_ bad) really sucks on an embedded device that has been stripped down to its bare-bones to reduce the cost because the owners want to sell the business to some conglomerate ASAP. We couldn't even share resources with the opaque third party app so we couldn't use SqlCe ourselves because it blew the memory budget to have two connections open and we couldn't use Sqlite because of old visual basic architects that are stupid and are terrified of anything that isn't microsoft. Also none of that matters because the architecture was so insane anyway and was never going to work because the devs didn't understand the limitations of the Compact Framework and used XmlDocument as a key part of the architecture. One transition from one screen to another would load (at a minimum) two very large Xml files and the Compact Framework and the hardware just wasn't good enough to do that all the time. The entire project was built around the hope that the hardware and framework could lift a lot more than it could in practice. As an aside, I was brought in to help fix some of the debris and work on the next project (which they were going to do before the Melbourne money showed up). This was after they finally removed the "must have a degree" requirement which had bounced me the years previous. Having a degree or not wasn't the problem, the problem was the devs they employed didn't have enough embedded experience and they (like most people at the time) drastically underestimated the complexity of embedded as well as overestimated how effective the compact framework was (it kinda sucked ass tbh). Thankfully most of this doesn't matter anymore because the company I'm talking about got sold to some big firm who lost the market to an up and comer some decade later.