Nice, I built a radix version of this in 2013. It didn't have a GC though (was left for later :D)
Mine was networked as an overlay to a linearizable range scannable kv-store with single-writer multiple-reader access and a three party model with the data owner just periodically signs the root to continue to delegate trust.
Had n-ary nodes though, the nice thing is that the proof is equal to the path to the node in that case (well because of radix/trie).
Was always playing with the idea of writing it down especially with all the crypto bullshit going on, but I was too lazy.
And merkle trees are what give "blockchains" their tamper protection: each state has a hash which includes the previous state's hash. (Although it'd be a list instead of a tree in the BC case.)
What's the use? I imagine 95% of companies that pretend to need blockchains could use something like this instead.
Something like off chain calculations or state management which then can be subtree’d out and used on chain. I think something like 1inch or Paraswap style of app were you need to do huge, in single tx terms, calculations to optimize trades across dexs and then provably import that work on chain. In reality because this is so new I doubt many people have any idea how to actually use it effectively right now but there is a lot of potential.
I’ve been playing with https://github.com/codenotary/immudb too. Very active community and the maintainers have been super responsive on their discord.
Had n-ary nodes though, the nice thing is that the proof is equal to the path to the node in that case (well because of radix/trie).
Was always playing with the idea of writing it down especially with all the crypto bullshit going on, but I was too lazy.