Follow

really i just want livejournal with a p2p data layer and an ecosystem of weird UIs

thinking about decentralized social media 

ok so i have been thinking lately that you can do better than just getting other clients to respect an author-initiated delete. you can disavow data and put anyone still peering that data on notice, helping to identify bugs in clients ("hey friend i noticed your peer hasn't dropped from that thing i deleted") and bad actors. in http-land that's hard to do, but it's simple when all your data lives in torrents linking to torrents.

re: thinking about decentralized social media 

more than create a p2p social media network, i think you need to take moderation primitives all the way to the data layer. i've been premeditating an architecture to support allow/ban-listing peers and dis/avowing records for the hypercore protocol, probably as a pouchdb database. see, i'm technically not writing a database!!

re: thinking about decentralized social media 

pouchdb plugin***

re: thinking about decentralized social media 

i can't tell you how much i wish there was a rust impl of hypercore. free me from javascript, someone, i beg of you

re: thinking about decentralized social media 

anyway, i will publish a readme before i publish code. i want to know what the fuck i'm doing before i do it.

re: thinking about decentralized social media 

@garbados "technically" is such a useful word

thinking about decentralized social media 

@garbados that's actually a really cool idea which I haven't heard proposed! I'm going to be thinking about this one a lot

thinking about decentralized social media 

@darius @garbados
That "it's easier to secure consent for deletes when your content is part of a distributed data store than it is when everyone keeps a separate copy" is part of the idea behind #datashards

CC: @cwebber @emacsen

thinking about decentralized social media 

@yaaps @garbados @cwebber @emacsen I'm afraid I don't see how that part in quotes relates to what I was commenting on? I'm not saying you're wrong here I'm just saying I don't understand how a distributed data store is specifically relevant here, and it might be, but I don't get it

thinking about decentralized social media 

@darius
This is what I thought you were replying to:

toot.cat/@garbados/10496257930

"you can do better than just getting other clients to respect an author-initiated delete... in http-land that's hard to do, but it's simple when all your data lives in torrents linking to torrents."

Is it the threading or is it that much of a stretch to compare torrents to something described as Tahoe-like?

@garbados @cwebber @emacsen

thinking about decentralized social media 

@yaaps @garbados @cwebber @emacsen you are correct about what we are discussing! I think I am just not understanding some terms of art here

thinking about decentralized social media 

@yaaps @garbados @cwebber @emacsen (I am not very good at understanding decentralization technical jargon and usually need things whiteboarded for me one on one before I get them, aka I am a bit slow)

thinking about decentralized social media 

@darius @garbados @cwebber @emacsen
Torrents are usually thought about in terms of distribution rather than storage, but the main difference in many cases is what the community expects for seeding

I'm not deeply literate in torrents, Tahoe, I2P, or Dat and there's obviously other points of difference, so you had me worried about whether I was actually being relevant

re: thinking about decentralized social media 

@yaaps @darius @cwebber @emacsen ok so i don't know what Tahoe is but i looked up DataShards and it seems to be this thing:

datashards.net/

> Datashards brings secure storage foundations to the modern web. It allows users and application developers to reason about secure, private data online or offline.

Sounds similar. I see OCAP, which I recognize but don't understand.

> Datashards is still in early development and is subject to change as it becomes more mature. As of now there are no official, formal specifications.

Short of a spec, there is a rationale:

github.com/WebOfTrustInfo/rwot

i have no idea what it is or how it relates but it seems interesting so i'm reading into it.

re: thinking about decentralized social media 

@yaaps @darius

oh, there's a python implementation:

gitlab.com/emacsen/pydatashard

this seems like a deep pre-alpha project, but it is concerned with the same essential problem of providing decentralized privacy and security assurances at the data layer. i'll have to read more; it's relevant prior art (but? not prior? living and contemporary?)

re: thinking about decentralized social media 

@garbados
Living and contemporary. Moving at the speed of free software 👍

re: thinking about decentralized social media 

@darius yeah i started thinking about it when i read the source on how cabal peers do backchannel comms to do handshakes, using a hypercore archive like a meeting point. in the same way an author-peer, having disavowed an archive, can instruct other peers to stop peering / drop cluster or be considered noncompliant. i need to write up the architecture, it's a doozy to try and fit in my head.

thinking about decentralized social media 

@garbados Potemkin users as ecosystem canaries.

thinking about decentralized social media 

@garbados oh that's a really interesting idea about deletes in distsys! New one, too, I think (though I'm not up to date on the literature).

Sign in to participate in the conversation
Toot.Cat

On the internet, everyone knows you're a cat — and that's totally okay.