Content
@
https://opensea.io/collection/neynar-1
0 reply
0 recast
0 reaction
artlu 🎩
@artlu
strong opinion, weakly held: client devs should not run Hubs (they should run a Replicator) - Hubs are expensive - point Replicator to nemes, StandardCrypto, Neynar - Replicator is copy of all (verified) data, without overhead of CRDT, syncing, network spam - SQL is more flexible+expressive than Hub HTTP API or gRPC
1 reply
0 recast
2 reactions
artlu 🎩
@artlu
@alexgrover.eth @haardikkk and @rish comments have convinced me to turn my perspective 180 degrees
1 reply
0 recast
1 reaction
artlu 🎩
@artlu
tagging @vrypan.eth for more obs about Hubs utility. Reading from Replicator, writing to Hubs may result in unavoidable async nastiness
1 reply
0 recast
0 reaction
vrypan |--o--|
@vrypan.eth
1. It's ok, if you don't mind about censorship resistance and self-sovereignty. 2. Replicator is ok for reads, but you will need one more integration to change the state. 3. "Async nastiness" is the beauty of decentralized hubs: at any given point in time, most hubs are not 100% in sync, but they eventually converge.
2 replies
0 recast
0 reaction
vrypan |--o--|
@vrypan.eth
I rushed to reply, and I have to clarify: If you have a fc client that requires me to install a backend to run it on my own, that's fine with me. So, if I need a replicator, or even your own custom backend, great, as long as I can install it on my own server.
1 reply
0 recast
0 reaction
artlu 🎩
@artlu
1. if we have a robust protocol, 2-3 dedicated orgs, >20 live Hubs, users get to free-ride on the decentralization, no? (a real Question!) 2,3. the async nastiness is mostly in UX (imo), users are used to TG/FB/X/Discord being near-instantaneous sync at 100000x scale
1 reply
0 recast
0 reaction
