I haven't read the whole paper yet so I might be missing the reason for this, but I'm surprised that the most significant P2P social networks I'm familiar with and actually use aren't included: Patchwork & Firechat. After a quick scan of the paper, the ones included have some various attributes that are presumably interesting from an academic research point of view, but I'm naturally curious to hear an academic's view of the above more actually useful (personal value judgment) apps, especially in comparison with the ones they included ....
Yeah, it seems like they're using "P2P" to mean "federated", which is backward. I'm surprised that they didn't think to include Secure Scuttlebutt in their paper.
Disclaimer: I work on Patchwork and SSB in general. :~)
Yup, SSB is a major one missing. Admittedly I was looking for my own network, Aether, which has more users and nodes than some of those listed — but if SSB is missing, then all bets are off.
I've honestly never heard of most of the things in that list and I spend a lot of time in P2P circles. Interesting.
Aether has such a compelling pitch that I was going to install it despite the Electron bloat it'd introduce on my system, but then saw that you used snaps, which was really unfortunate.
Have you considered something like AppImage instead? While not ideal, it minimizes your effort while also being compatible with every other distribution without requiring your users to install a large, complicated piece of software/package manager like Snap.
Aether is a Go app, we use Electron just for the GUI. We do no processing in it - we do the processing in the go backend daemon. It’s pretty snappy and light.
If you're interested, I think it would be really wise to have our DIY P2P communities chat more often. It's great to all get together at events like Dweb camp, but since we're all working on similar problems maybe more communication is better?
Yes! I’m chatting semi-regularly with a bunch of SSB people over Twitter, but I would love to chat / help out more. SSB is a much bigger project than Aether so I’m sometimes at a loss who to talk to since I don’t know the project structure or who the core developers for certain features are. My email is in my HN profile, please do reach out.
Is anyone here a fan of Darius Kazemi? He's one of the people who I believe has a pulse on the future of the internet. He has a great document on how to roll your own federated social network:
If you're talking about TimBL's Solid, I wouldn't call that a social network: it's a set of specifications for data storage, on top of which people could build a federated social network.
(Disclosure: I work at inrupt, TimBL's startup, but this is my personal opinion.)
What's the difference? Is it that P2P lacks a central server but federated is distributed servers (i.e. no single central server, just a bunch that can authenticate against each other)?
Isn't GUN the one with the lexicographical LWW? It may allow peers to communicate between each other, but I would fit separately from WebTorrent or SSB.
