Hiroshi Lockheimer once confirmed that Fuchsia at this stage (to be exact, 4 years ago) is more of testbed for OS technologies that cannot be readily integrated into Android. It is quite absurd to say that Fuchsia is a competitor against Android. It is more close to Midori with a slightly clearer path to productization.
Yeah, I think that's the ambitious moonshot at least for Fuchsia team, and Google might hope it to be realized. But it probably also acknowledges that it's a very unrealistic goal. More likely scenario is to gradually replace some important core systems (including its kernel?) with Fuchsia while keeping the overall Android ecosystem.
I agree. I don't think they can abandon the Android ecosystem at all, I mean all the apps and the store, that's the real value. They could design a migration path for the apps and make them work seamlessly on Fuchsia in the meantime, gradually replacing the ecosystem with the promise of new "shiny" features for developers. Building an OS from scratch is very expensive in terms of resources and money. I cannot see a valid technical reason for this move. My view is that many projects are now using AOSP to build their own operating system and trying to get rid of Google services, which is a threat to Google and its business based on bombarding the user with ads. The biggest obstacle at the moment is getting the applications that rely on Google services to work. e/OS/ uses MicroG and in my personal experience everything works seamlessly, including banking applications. The other crucial aspect is the availability of stores for the apps. Aurora is just an alternative front-end client to Google PlayStore, but it is a huge step forward in removing direct dependency. e/OS/ has AppLounge which does the same thing.
I'd certainly prefer to download and install my bank's app directly from a protected area of the bank's website rather than from a generic store. Implementing a custom authentication mechanism (e.g. signed with GPG) and an auto-update feature is certainly doable.
