Rasp Pi model 3 WiFi

So basically a WINE type approach? Yeah, that’s a lot of work. And I think it would effectively mark the death of the OS – development for the native version would stall (there’s no point if the virtualised version offers multi-threading, hardware-accelerated media playback, modern networking, etc.), and over time the APIs which the virtualised version exposes to the programs would mutate closer and closer to those of the host OS until there’s nothing left except a bit of RISC OS themed UI fluff.

Of course if users only care about the applications then the death of the OS doesn’t really matter. But I’m more interested in the OS than the apps, so it’s not an approach that appeals to me.

One of the reasons why I’m not convinced that a WINE type approach is even worth it is that you end up supporting a load of APIs that aren’t really that great to begin with. You need them to be there for compatibility with older applications, and you don’t really want to encourage new applications to use them. Ultimately, if you want to support the legacy APIs, you might as well run old applications in emulators on faster hardware or choose a virtualisation solution that lets you reuse much of the (old) operating system in a sandbox.

If you evolve the OS, you may end up with a set of APIs that are incompatible with the ones that Acorn left behind. If Acorn had done what Apple did with the old Mac OS, there’s a chance that much of the existing APIs would have been deprecated anyway.