Dan Romero pfp
Dan Romero
@dwr.eth
Holiday week AMA! What's on your mind?
43 replies
0 recast
0 reaction

Henry de Valence pfp
Henry de Valence
@hdevalence
is it possible to build a farcaster hub that only maintains a subset of the network state? if so, is it possible to avoid having a distinction between “clients” and “hubs”?
2 replies
0 recast
0 reaction

Varun Srinivasan pfp
Varun Srinivasan
@v
sharding state affects availability of data by making sync harder and making it easier to censor people you don't like. we're actively designing hubs to ignore peers that don't relay the entire state of the network
1 reply
0 recast
0 reaction

Henry de Valence pfp
Henry de Valence
@hdevalence
i don’t agree, i think this is only true without an end-to-end conception of the protocol design that excludes the devices people actually use from consideration… thereby forcing users to rely on hubs and making censorship easier in the long run
2 replies
0 recast
0 reaction

Henry de Valence pfp
Henry de Valence
@hdevalence
(blanket disclaimer that, like, it’s your protocol, not mine, so feel free to discount my bystander opinions to whatever extent you want)
0 reply
0 recast
0 reaction

Varun Srinivasan pfp
Varun Srinivasan
@v
i'm open to changing my mind if you have a proposal for how to make sharding work while protecting users against eclipse attacks and keeping sync performant
1 reply
0 recast
0 reaction