I’ve become a relatively big fan of resurrection; I like the idea of skip proofs at a 1 week granularity, and the min-1-week-TTL was the missing component that I did not understand in the “proofs of non-existence” schemes. So super ACK on this.
A few questions:
- Why keep the reasoning at a “contract” level at all? Why not just have storage keys, which users pay rent on, and these storage keys can happen to store contract code? It’s purely a terminology difference, but this seems cleaner to me, and unifies contract code and data storage without requiring all data to be stored in “contracts”, which makes less intuitive sense for discussion, IMO. Of course you can say that these “contracts” are basically SLA contracts on storage locations, but I think this is sort of implicit in a rentful model when talking about data. The lack of distinction between “data” and “code” and the idea of executing what is stored at a given address in “memory/state” is also highly intuitive from classic programming paradigms.
- Is there going to be a small incentive for poking? Should there be one?
Personally I favor the “landlord / storage key” terminology.