We do something similar across a React and a Next JS app. A little different because they are both web, but same principles of letting the clients do their own thing and having common backend logic.
Also just started looking around your site. This is actually a quality engineering blog! We're actively implementing stylized components in a manner similar to yours.
I guess it's "from the ground up" as in "greenfield", where there's no preexisting code to support or rewrite. I.e. "ground" means "fresh project" (and not e.g. entering bits by hand).
We considered Flutter, React Native, Kotlin Multiplatform Mobile and pure native as solutions. The factors we looked at were size of hiring market, ease of conversion of web codebase and the effectiveness of the current team in that framework.
This is one of the most pathetic answers possible. You give nothing on user bases and adoption of both dart and flutter, internal adoption within google, it's use of Dart and Flutter for the frontend of Fuschia, googles desire to control a widespread language, the open-source nature of Flutter, the list goes on. You just provide a list of PRODUCT's that have been killed off, not an entire language and frameworks. Your idea that google will just kill it off is a meme, and doesn't take into account it's usefulness to google in the wider world.
In terms of user bases, Orkut (RIP) had north of 60 million users when Google killed it. It was the primary social network in multiple countries.
Maybe the biggest user base killed off was Google Reader (RIP), it was the world's foremost RSS aggregator with no forseable competition and no obvious exact replacement for all those users to go to when GReader was turned off.
In terms of languages and frameworks, the Noop language was killed more by Kotlin than by Google but Google App Engine has been crushed even if the name lives on, the old SDKs you relied on are mostly gone. There all alternatives, but not quite as attractive.
Lastly though, that page i linked has a LOT of examples on it. I feel that adds weight.
Hopefully this longer comment goes some way to assuage the idea that my prior comment was a thoughtless throwaway.
For those who have the pleasure and luxury to go green field new dev, i started using the Quasar Framework, which builds on top of Vue. I really love it.
> Quasar’s motto is: write code once and simultaneously deploy it as a website, a Mobile App and/or an Electron App. Yes, one codebase for all of them (...)
> the opportunity to approach the challenge of supporting multiple platforms without the constraints of legacy engineering decisions
While I am keen on this article, I must pause reading it to come here and plead that people say "prior" instead of "legacy".
Legacy - due to its provenance in Microsoft sales techniques to mean "not Microsoft" - implies "bad". When actually the decisions may have been good, at least at the time.
> On mobile, we generate a QR code that can be scanned by a phone running the Ambrook app to download and run the latest Javascript bundle live on the phone.
That's so important when you're shipping at startup speed and it's so tempting to just rush to the next thing. I am going to try this with my next cross platform app.
Of course, why wouldn't they use the latest stuff? Farming technology is literally the reason you and your (relatively) recent ancestors aren't slaving away in the fields. Doesn't matter if that's combine harvesters or logistics apps.
Also just started looking around your site. This is actually a quality engineering blog! We're actively implementing stylized components in a manner similar to yours.