[pjmlp]

Including desktop development, as one ex-advocate for WinRT, which I considered to be what .NET 1.0 should have been in first place, given the original idea of COM+ Runtime vNext, they really messed up the whole development story multiple times, always demanding a rewrite as if that is a minor issue.

However, nowadays, run away from anything that resembles WinRT tooling, keep to tried and true Win32, Forms, WPF, and even MFC, or plain Web, versus anything that builds on top of WinRT, unless the desired API is only available via WinRT.

[gwervc]

> WinRT

It was obvious from day one that this would be a failure since it was strongly tied to the Metro UI (awesome on mobile, ugly as hell on desktop) and the Windows app store, all only available on an OS that never was competitive against its predecessor. Also it shipped with a .net implementation that was incompatible with the real one. I'm glad I skipped that tech entirely.

[pjmlp]

It evolved over time, by the time Project Reunion was announced in 2020, it had gotten kind of ok.

However, that incompatibility you point out, was present throughout all iterations.

Window 8 => 8.1 => 10 required rewrites, UWP/WinUI 2.0 => Win32/WinUI 3.0 dropped .NET Native and .NET 9 still isn't full AOT, C++/CX => C++/WinRT lost Visual Studio tooling and is nowadays in maintenance mode, yet gets sold as being the way for C++ devs.

Meanwhile, most of the faces on community meetings have changed since Windows 8 days, nowadays most seem fresh out of university without any Windows developer experience given the blank stares when questioned about feature XYZ becoming available, and bugs?, you only have to spend sometime digging around the Github repos.

Naturally, even the strongest advocate eventually gives up.