I was building a touchscreen music-production app for Windows, and I wanted it to have a phone/tablet style interface. It would have a stack of modal popup Views that the user could tap (or click) off of to close the top View and go back down. I ended up splitting the UI system off into a separate library and kept working on it. It now has DI (with scoping) and some other features. I've since used it for a few other small personal tools and projects, though no one else is that I'm aware of.
Also question to the room, is DI really worth it ? All that complexity to switch out sub-classes . A clean design but in practice it just feels like too much work for the benefit.
Thanks. The music app code is closed. There's a public demo still online, but I got terrible feedback, so I'm slowly rebuilding it from scratch with a completely different design.
As for DI: in my case, it was definitely worth adding. The original music app didn't have it, so I had to come up with convoluted mechanisms to share cross-cutting services with different ViewModels, and it got messy. Some had to keep references to services that they didn't actually need themselves, but their child ViewModels did... With DI, I just register the services with the library, and let it inject whatever a ViewModel needs.
It's written in C# and runs on top of WPF: https://tinyurl.com/upbeatui