Secure Scuttlebutt is a really cool system. At the bottom is a messaging system, and on top of that people build specific tools. For me the killer app is git-ssb.
I was on gitorious.org once upon a time. Then that system got bought and I moved to gitlab.com. When will that get bought? I should set up my own instance, maybe, but who has the time? Setting up git-ssb is trivial, just an npm install away. Now I can push my commits to SSB. A fellow hacker on a project can access it over their local git-ssb instance, and if you run git-ssb-web you can read and write issues. And it works when you’re offline too.
For the general public, you can give them a link to one of the public gateway sites. If they go down, you still have your data, and the network has your data, the web frontends are just a convenience. Anyone can put up a web frontend.
Having said that, anyone starting with ssb should go to patchwork, the social network, first, because the instructions for setting up that include getting started with ssb itself, and patchwork also has a UI for doing things like choosing your name and making the initial connection to the network.
I will repeat the author’s parting words and say that you should check out the documentation and give it a try.
With typical gossip, however, information deteriorates as it spreads and eventually becomes harmful rumor. Scuttlebutt on the other hand makes word of mouth secure with cryptography.
I can’t say I’ve ever seen this happen, at least not in any way that crypto would solve it. If something gets retweeted 10000 times, it’s just as truthy as it was to start despite the lack of crypto signature. Nor does throwing crypto on top of a rumor make it accurate.
So how do nodes actually find each other?
Addresses of pub servers are shared on the network and used for connecting. Nodes on the same LAN discover eachother using UDP broadcast packets. Using an invite code triggers connecting to a node at a specific address.
This is neat. I’ve been wanting to build a “personal internet” for a while. Basically the idea is that you could own and host your own data, and if you wanted to share that data it would get sent peer-to-peer across whatever network you were connected to (whether that was the Internet or some kind of mesh network). Obviously, you would also need to be the authority of your own data too, and (although I haven’t read this in detail) the solution offered here seems to solve that problem nicely.
is there specification somewhere ?
What happen if one fork it’s own log of event?
Peers won’t accept new messages that reuse a sequence number of an existing message in the feed. So they will stop being able to replicate that feed.
but why node.js ?
and have they considered rewriting it in rust /s
it’s a pun ?