The data is not centralized, but everyone is using the same aggravation service (indexer) to access the data.
Comment on What about AT protocol?
CanadaPlus@lemmy.sdf.org 1 month agoReally? In what way?
Digital identities being cryptographic and independent of any one instance is huge all on it’s own. The rest of it I understand less clearly, but it looks pretty modular.
ericjmorey@discuss.online 1 month ago
GregorGizeh@lemmy.zip 1 month ago
Don’t aggregate me pal
ericjmorey@discuss.online 1 month ago
LOL I should have reread that one.
cecilkorik@lemmy.ca 1 month ago
Most aggregation services are also aggravation services, so this really makes sense either way.
gandolfini_the_grey@lemm.ee 1 month ago
See here. Basically, creating a relay for the AT protocol is extremely costly and only possible for big tech companies with lots of investors.
CanadaPlus@lemmy.sdf.org 1 month ago
Thank you!
TL;DR, the relay bit works as a completely connected network topology, and has the associated quadratic growth issues, which renders it hard to host.
Also nasty: Direct messages are just not federated.
Other things are or were at the time of writing janky, but nothing else is quite that egregious. The author is working on a separate project, and recommends this idea as a solution for portable identity on ActivityPub, and here’s what “object capability” means.