This is hardly true. It's easy to keep a history of query logs, which is enough to figure out 90% of data lineage. The rest 10% probably correspond to operations not even runnable on a blockchain.
I can mock a web3 application by opening a database to the entire internet. All edits logs will be visible, although edits will be access controlled. Now I'll claim this is observable. Some people would probably complain. But how in the world is the general public going to complain?
I still think it could be potentially useful to, with a centralized database, use a similar kind of authentication scheme, in a way that allows for succinct demonstrations of any kinds of rule violations (compared to the rules that were set), as well as efficient to check proofs that state changes are in accordance with the rules which were set.
(I wonder if a network of such centralized DBs run by different parties might be enough to produce incentives among the different DB owners to behave in accordance to rules which they initially publish for themselves, etc. )
This is hardly true. It's easy to keep a history of query logs, which is enough to figure out 90% of data lineage. The rest 10% probably correspond to operations not even runnable on a blockchain.
I can mock a web3 application by opening a database to the entire internet. All edits logs will be visible, although edits will be access controlled. Now I'll claim this is observable. Some people would probably complain. But how in the world is the general public going to complain?