my kingdom for an actually functioning first-class social web (and not a bunch of silo clones co-opting the term)
-
the *real* missing building block is an identity that is consistently recognized everywhere across the web
couple that with a standard way to publish web resources as that identity, and you have basically everything you need
@trwnh IMO the Nostr approach of keeping keys in the client and having relays be more simple is good. Key management and multi devicw identity is a major PITA, but there's a lot of known approaches that can be taken.
-
the *real* missing building block is an identity that is consistently recognized everywhere across the web
couple that with a standard way to publish web resources as that identity, and you have basically everything you need
if you're willing to accept DNS in all its faults and limitations, then we could devise a system where every social agent has its own FQDN. heck, that's basically bluesky's facade. the rest of the system mostly falls into place after that -- just make all links relative to the social agent's FQDN, and keep track of when the FQDN changes. https: can do this with some maintenance (rewriting old links), did:whatever can do it more dynamically with service + relativeRef
-
if you're willing to accept DNS in all its faults and limitations, then we could devise a system where every social agent has its own FQDN. heck, that's basically bluesky's facade. the rest of the system mostly falls into place after that -- just make all links relative to the social agent's FQDN, and keep track of when the FQDN changes. https: can do this with some maintenance (rewriting old links), did:whatever can do it more dynamically with service + relativeRef
honestly you could just maintain a petname / contact-or-address-book type thing using redirects because HTTP lets you do that; you don't need to pre-canonicalize all your links
say my http server redirects /proxy-to-my-friend-bob/foo -> bob.example/foo
-
@trwnh IMO the Nostr approach of keeping keys in the client and having relays be more simple is good. Key management and multi devicw identity is a major PITA, but there's a lot of known approaches that can be taken.
@mauve it's already pretty widely accepted on the web that DNS mostly works, so we can also establish identity that way (say we agreed to give everyone a FQDN so they can do TLS, or we adopted the CID spec https://www.w3.org/TR/cid-1.0/ so that they could use some URI or DID). aside from that we could also distribute bearer tokens out-of-band, but we have better options
-
my kingdom for an actually functioning first-class social web (and not a bunch of silo clones co-opting the term)
@trwnh@mastodon.social Build it
-
@trwnh@mastodon.social Build it
@natalie that's the aim and intent. the hard part is getting everyone to agree on a federated identity system (and to a lesser extent, a standard API for publishing web resources using that identity)
-
@natalie that's the aim and intent. the hard part is getting everyone to agree on a federated identity system (and to a lesser extent, a standard API for publishing web resources using that identity)
@trwnh@mastodon.social everyone agreeing is an unattainable goal
-
@trwnh@mastodon.social everyone agreeing is an unattainable goal
@natalie to some extent we all converged on using http(s). i think we can do something similar again, or at least i would hope so
-
honestly you could just maintain a petname / contact-or-address-book type thing using redirects because HTTP lets you do that; you don't need to pre-canonicalize all your links
say my http server redirects /proxy-to-my-friend-bob/foo -> bob.example/foo
i think you could get pretty far with existing technologies like webdav or solid, but i'd really want to see a better way to manage "name to thing" resolution that isn't literally writing an nginx config with location blocks
-
@natalie to some extent we all converged on using http(s). i think we can do something similar again, or at least i would hope so
@trwnh@mastodon.social converging is much different than "getting everyone to agree" tho
-
@trwnh@mastodon.social converging is much different than "getting everyone to agree" tho
@natalie admittedly that was poor phrasing, depending on the definitions of "everyone" and "agree"
in a practical sense i think the closest current option is to assign everyone an FQDN, since that requires the least changes
-
@natalie admittedly that was poor phrasing, depending on the definitions of "everyone" and "agree"
in a practical sense i think the closest current option is to assign everyone an FQDN, since that requires the least changes
@natalie i'm still trying to come up with a better answer but that's the societal baseline at least
-
@natalie i'm still trying to come up with a better answer but that's the societal baseline at least
@trwnh@mastodon.social bsky wins again
-
@trwnh@mastodon.social bsky wins again
@natalie yeah i don’t fully agree with all their decisions but i understand why they made them
virtual host support would probably be quite feasible to retrofit onto existing fedi codebases but i doubt it’ll get done anytime soon