hypothesis: if too many early users sign up on mastodon.social instead of other instances, federation will end up becoming a second-class feature. (imagine what email today would look like if gmail had existed in the 90s.)
other instances need to be able to compete with mastodon.social for growth, or https://github.com/tootsuite/mastodon/issues/177 should be implemented.
@bcrypt Federation will never be 2nd class feature for me.
@Gargron @bcrypt I think that "too many users" is subjective (apart from scaling). If anyone thinks there are too many users around here, or too many douches, they can either move one, or run their own instance. This way, putting more pressure on any system, could help to balance the power over the federation as a whole.
Thoughts?
@tim @Gargron i am reminded of https://mako.cc/copyrighteous/google-has-most-of-my-email-because-it-has-all-of-yours as an example of too-many-users-on-one-node-of-a-supposedly-federated-system
@sarahjeong @Gargron @bcrypt Working on it.
@tim @sarahjeong @Gargron @bcrypt Wait, working on it how?
@tim Oh wow, cool. I wish I could afford to quit my day job and work on interesting projects.
Are you actively attempting to integrate this into Mastodon atm, or just talking long-term?
@kevin_redacted Working on a proposal that includes self-issuing identities, managing multiple identities, accurate tracking of accompanied reputation, controlling one’s social network(s), sharing sensitive data and/or personally identifiable information (PII), trolls, sock puppets, and Sybil swarms.
Give me a few more weeks and I'll drop it here first. :)
@kevin_redacted Oh, and there's also gonna be multi-sig, distributed (read: mobile) reputation, some P2P networking, strong crypto (for bothing signing messages and encrypting data), local data storage, and user- and/or machine-controlled sharing of (sensitive) data.
Should address heterogeneity, non-centralisation, micro-intermediation, pubkey namespace, WoT, JSON msgs, APIs, CLI, revocations, Sybil protection, DoS protection, etc.
@tim Sounds like you've thought things out a BIT more in depth than the kind of quick hack sort of solutions I was contemplating, lol
@kevin_redacted Since ~July 2014. It has been crystallizing for a while. I already can't wait to see how I'll get sent back to the drawing board. I've already proposed it to a few folks, and haven't found any major flaws so far. Smartest kid in the room principle though.
@kevin_redacted Right now I'm just super-excited because so many other people start thinking about, and discussing, the subject. The amount of fresh point-of-views on it is amazing. :D
@tim I bet
@kevin_redacted Prepare the reading glasses. Buffer overflow incoming.
@tim @kevin_redacted That sounds really cool. Do you happen to have anything written down about the technical design? I would love to read more about it.
@cdata @kevin_redacted It's heavily inspired by Identifi, and am currently working on a proposal with many more details. I'll dropping it here first. :)
@cdata @kevin_redacted *drop. And here's an article by Identifi's creator, which may interest you: https://medium.com/@mmalmi/learning-to-trust-strangers-167b652a654f#.2eyijxngr
@kevin_redacted @sarahjeong @Gargron @bcrypt
Quit my day-job as an IT manager 2 years ago to work on the topic of P2P identity & reputation. To get an idea, see: https://github.com/identifi/identifi