yan 🐇 is a user on mastodon.social. You can follow them or interact with them if you have an account anywhere in the fediverse. If you don't, you can sign up here.
yan 🐇 @bcrypt

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 github.com/tootsuite/mastodon/ should be implemented.

· Web · 89 · 117

@bcrypt A nice feature of Mastodon is that you can set up a single-user (aka no-registration) install. Which makes it perfect for running your own federated instance.

@bcrypt Federation will never be 2nd class feature for me.

@Gargron @bcrypt Even financially, if anything, if no one else builds something competitive in fedispace, Eugen is incentivized to build parallel instances tailoring to a few different community styles, each with their own Patreon. I'd bet this would be more profitable than forcing everyone to stay on M.soc

@five @Gargron @bcrypt I'm so excited about all of this. Watching all of this unfolds is amazing by itself. Considering where all of this could go is ... maly.io/media/7jyt-RTuT-wUMRVd

@five I think the "community" of instances is kinda overrated. I don't identify as part of the "maly" community. I think all 64 of us are mostly just here because the admin is cool and it's probably going to stick around for awhile

@five nevermind there is like 103 of us now. where did everyone come from heh

@kodo my best guess is it'll play out like subreddits: there's a Reddit culture that spills into everywhere, but the biggest subs also have some cultural artifacts of their own. Then smaller subs will be delineated based on niche uses and rulesets and whatnot

@kodo the real-world analog I see is the school cafeteria. Some people rigidly stay in their one table; others float among several. Some are mostly loners, with a really small circle or maybe no one else at all. Anyone can try to join any table but if they don't fit in, the table will eject them, either explicitly or by snubbing them

@kodo shit and I guess for some rare personalities, tables may deliberately court them

@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 mako.cc/copyrighteous/google-h as an example of too-many-users-on-one-node-of-a-supposedly-federated-system

@bcrypt @tim I have closed down registration on this instance to prevent this

@Gargron @tim @bcrypt
Hopefully symbolism like this will help pre-bake the culture into one that might avoid that trap.

I feel like this is a race to see who can run the largest instance so they can get advertising $$$ somehow :(

If we can get something like the Patreon model that @five talked about it might sidestep the incentive feedback loop further. I wonder if there are any easy to use decentralized server providers that can get fed money from shared cryptocoin wallet.

@ultimape @five @bcrypt @tim @Gargron I think a instance that's free to users with moderate and transparent ads could be a good thing. Particularly if it kicked revenue back to users based on content creation? So people could choose to pay a subscription to a node for ad-free experience, or transparently participate in a free to use space ad-lite that comps users any revenues beyond operating costs.

@mattcropp I personally don't think that would scale well. Hard to get advertising money without strong deals, and I think ti would only encourage users to game the system.

But I might be biased. I'm still not done writing my anti-advertising thing and it's 10k words (including quotes).

@ultimape @Gargron @bcrypt @five How about a multi-sig wallet, like on BTC.com? Easy to user, and still usable without the third party.

@Gargron @tim i feel that was the right choice. i would like to run an instance but don't want the burden of maintaining it forever; giving users a migration strategy is a deal breaker for me running one

@bcrypt @tim @Gargron for me I would like to see a way to port follower-lists to other instances, I don't know if that's feasible but it would help being able to move and prevent supremacy of any one node

@sarahjeong @Gargron @bcrypt
it also makes it more resilient if an instance dies
you can tell your followers to refollow you, but realistically that won't happen unless you are wicked popular

@bcrypt Late to the convo, but that's exactly what happened the first time OStatus went big-ish. Everything centralized around status.net, then identi.ca, which then turned into pump.io when Evan wanted to make his New And Improved thing (ironically just as W3C WG got underway iirc).

@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: github.com/identifi/identifi

@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

@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. :)

github.com/identifi/identifi

@sarahjeong @Gargron @tim @bcrypt

I agree. Some sort of way to keep the userbase mobile is going to need to become very important.

@knowtheory @rich_harris @bcrypt @Gargron @sarahjeong

Issues were next on my to-do list, to study. Thanks for the shortcut! :)

@knowtheory @sarahjeong @gargron @bcrypt @tim @rich_harris ah word, the "account redirect" seems like it would solve a lot of problems.

@schlink @rich_harris @tim @bcrypt @Gargron @sarahjeong Heh. There are a lot of distributed computing problems hiding in here. I haven't caught up w/ the issues well enough to have an opinion about how scared i should be yet :D

@tim @schlink @rich_harris @bcrypt @Gargron @sarahjeong Wait, is this saying that the CAP theorem is newspeak? :D

@bcrypt Are you saying that you won't run one if user's can migrate away or the opposite?

@bcrypt Why is a migration strategy a deal breaker?

@cr1901 see the part of the sentence before the semicolon

@bcrypt I think I'm missing something obvious. Wouldn't giving users a migration strategy mean that you *don't* intend to run an instance forever? Or is a migration strategy a recurring burden as long as you run the instance.

@Gargron @tim @bcrypt Maybe throttle, instead of close down? I suspect it's an important time, and total growth may be as important as federation-balance? I really want a Twitter replacement so I might just be too excited...

@Gargron [jumping in] i appreciate the thoughtfulness you're putting into community building here [jumps back out]

@bcrypt @Gargron @tim does mastodon deliver all messages to all instances?

@bcrypt @Gargron Yep. Centralisation is efficient. Still, lots of different power dynamics here that we can study. #crowdpsychology

I expect the #Fediverse to break up into many smaller pieces, as I've elaborated on previously: medium.com/@2W/there-will-be-m

@bcrypt that would seem to only affect maybe Unlisted posts, no? (i.e. Public is already expected to be global; private and locked limits scope as intended. This assumes proper encryption of course.) @Gargron @tim

@bcrypt @Gargron @tim which is not even to mention identi.ca, which is literally the predecessor to GNUsocial/mastodon!

@bcrypt @Gargron @tim power law distrubutions get you every time

@Gargron @bcrypt fam the flagship node blocks a significant amount of the network. pls

@Gargron @bcrypt I also think, mastodon.social is now first class, and all other servers are 2nd class. Also there is not a clear 2nd choice available (as new user, which server should i choose? Where are the infosec people on?).

Imho a 2nd big server aka mastodon2.social would be ideal.