Zach
@zachterrell
@v Just wondering, if the L2 Delta Graph is robust enough to guarantee consistency for messages, why not just use it for the identity layer as well?
2 replies
0 recast
0 reaction
Dan Romero
@dwr.eth
How do you resolve who the real @zachterrell is?
3 replies
0 recast
0 reaction
Zach
@zachterrell
In the sense of consistency or actual real world identity? I assume you mean the former, in which couldn't a "registration" just be represented as a delta type the same way a "like" or "follow" is represented?
1 reply
0 recast
0 reaction
PFH
@pfh
The problem has to do with ordering. For registrations, we care about who was first. But for reactions (and most data on hubs), we care about the most recent state (who was last).
2 replies
0 recast
0 reaction
PFH
@pfh
A blockchain is good for the first case, because (in exchange for a performance hit) you can prove that something occurred at a specific time in the past.
1 reply
0 recast
0 reaction
PFH
@pfh
Farcaster CRDTs are good for the second case, because while we can't validate that something actually occurred when you say it did, hubs are very fast and can block messages that claim to be from the future.
0 reply
0 recast
0 reaction