Hello! thank you for posting this proposal. I read through your document and researched a bit about it.
According to what I could find online, JSI-Native support adoption on React native development is somewhere from 25% to 50% at most. I don’t agree with your argument that your solution is “prepared for the future”. I’d rather it’s prepared for the present. I hope I’m wrong and that people would adopt Zcash at any cost, but I don’t see that any Engineering Manager with good judgement would actually rush their codebase’s evolution towards a new React Native Architecture to just adopt Zcash on their App. I think your application would be way better if it offered support for the 100% of React Native developers or at least… 90% of them with a bridge solution and a JSI solution so that they can just adopt Zcash and worry about their stack evolution later.
Another question:
Your design says that you would interface librustzcash with JSI, but at the moment librustzcash crates don’t offer a sync engine for light clients. They might in the future after Zallet changes are coded back to zcash_client_backend but it’s not available now. The mobile SDKs have a mixed approach (sorry! our bad! it was another era) and the sync engine is spread accross both sides of the FFI, how would you address that?
Thanks!