Conversation

Notices

1. Embed this notice
   13 barn owls in a trenchcoat (hauntedowlbear@eldritch.cafe)'s status on Friday, 08-Nov-2024 15:55:46 JST 13 barn owls in a trenchcoat

   • nilesh

   @nilesh it's early and I'm skimming, but AT's approach sounds akin to the idea behind Solid pods - user controlled data store, interoperable protocol, monolithic services that those can be plugged into?

   • Embed this notice
     Open Risk (openrisk@mastodon.social)'s status on Friday, 08-Nov-2024 18:42:58 JST Open Risk

     in reply to

     • nilesh

     @HauntedOwlbear @nilesh

     To me its all still quite confusing. Activitypub seems quite a bit less prescriptive, a lot of important details are "left to implementation". The Mastodon server network is one such fully developed architecture but quite different designs are compliant, e.g., single account "instances" like gotosocial. Somebody even suggested that the entire activitypub pattern could be implemented over the atproto protocol etc.

   • Embed this notice
     13 barn owls in a trenchcoat (hauntedowlbear@eldritch.cafe)'s status on Friday, 08-Nov-2024 18:42:58 JST 13 barn owls in a trenchcoat

     in reply to

     • nilesh
     • Open Risk

     @openrisk

     I'm a genuine enthusiast of the One True User-Owned Datastore model, although I've found current models of that to be offputtingly obtuse in deployment terms
     (I mentioned Solid earlier - see https://solidproject.org/about) - let alone for laypersons.

     As someone currently operating three different ActivityPub servers (Masto, GTS, WriteFreely), the fact that user data is still siloed on a per-server basis is a weak point, as if the server goes, the user loses that data. (Which I can solve by running my own servers, but the data remains on its home server. This has both advantages - I like specialisation and compartmentalisation! - and disadvantages, in that siloed, non-transferrable data becomes particularly sensitive when you don't run your own servers.)

     In an ideal world, I'd like to see user-owned data stores that could be plugged into ActivityPub on a protocol level, although I suppose that's not something that's going to be on the AP roadmap as it's fully mature now, but instead something that falls under the Spritely Project's area of interest for the next generation of open, federated protocols.

     I don't think the AT Protocol successfully overcomes the user accessibility to data stores problem. (And I've no interest in touching Bluesky and its projects for a variety of reasons.)

     @nilesh

   • Embed this notice
     13 barn owls in a trenchcoat (hauntedowlbear@eldritch.cafe)'s status on Saturday, 09-Nov-2024 00:16:46 JST 13 barn owls in a trenchcoat

     in reply to

     • nilesh
     • Open Risk

     @nilesh
     Because I run a blog (WriteFreely), a small community Mastodon server, and an even small personal server for a specific audience (GTS).

     But that was more an indication that I'm familiar with the issue I'm detailing (the lack of data ownership and portability).

     While I wouldn't necessarily personally want to cross the streams in terms of having a data store for, say, my porn server being the same as the one for my vanilla social networking, blog and mainstream professional life, the notion of single user controlled data stores that are platform agnostic is one I'd like to see as a key part of the discussion when we speak about the future of decentralised social media (and other!) protocols.

     @openrisk

   • Embed this notice
     nilesh (nilesh@fosstodon.org)'s status on Saturday, 09-Nov-2024 00:16:47 JST nilesh

     in reply to

     • Open Risk

     @HauntedOwlbear @openrisk Why do you need 3 separate AP servers though? AP apps are now starting to support multiple object types like <Article> instead of only <Note>, but this was never a protocol limitation.