The idea is that every individual person has a container for their own data, and then platforms interact with that data using methods built on the Semantic Web.
I frankly don't think it has a hope in hell of succeeding, but I love the concept!
Usually today your data is fragmented across platforms (so damage is reduced) which have centralized authorities who can step in and fix bad actor issues.
> Anyone or anything that accesses data in a Solid Pod uses a unique ID, authenticated by a decentralized extension of OpenID Connect. Solid's access control system uses these IDs to determine whether a person or application has access to a resource in a Pod.
Of course, as a data owner, you could accidentally grant a bad actor access to your data, but presumably you can also revoke that access as well.
They need a way to handle situations when bad actors take over, because other solutions handle this with centralized authorities who step in and rectify the issue.
Are you talking about literally exploiting a bug and hacking the underlying service that is providing access to the pod?
In that case, it's a question of who owns and operates the pod. Solid is conceived as a set of standards that can be implemented by either individuals, or by companies on behalf of individuals. Think "data ownership as a service".
So you can still have centralized entities that implement the spec and provide support and other services for users, including dealing with security incidents.
If a "chunk" of data could be "sovereignized" by association with other "nearby" chunks of data (through some common signing event), one could "move" seamlessly from one chunk to the other, regardless of where it was being "served" from.
When I talk about "chunks" here, I'm talking about document boundaries and ownership/control of the data.
If we were in some sort of simulation, moving from my sphere of data to yours would look something like a probability distribution. By having signed contracts between the spheres, one could increase the probability that movement in a general direction resulted in access and loading of trusted data.
Because the chains start out empty, nobody "owns" these fictitious plots until they are established through preexisting trusted channels. Various channels as well. None of this bullshit NFT land grab crap that nobody can trust. Use NFTs, or even better, Lightning payments to establish the "authenticity" of the nearby data spheres.
Properties: * fully self-authenticating protocol independent of DNS and the TLS certificate authorities * self sovereign identity * you can migrate to a different host and keep all your data, and social graph, and old links continue to work. * You can authorize and run live mirrors of your data which help provide bandwidth * you only see things from or via people you follow
If you know of a homomorphic encryption scheme that would work for this and remain economically viable for other people to host, let me know. I simply don't think it's reasonable to expect that degree of privacy, much less on a Twitter clone. If you want privacy, use Matrix or OMEMO.
This includes Radicle, IPFS, Skynet, SSB, Arweave, Ethererum, and many more.
On the other side, censorship is also very important, nobody wants to have to dwell into spam and filth to read a small amount of valuable content. Maybe be able to choose your own censors somehow? Or build chains of trusts with your friends and "friends-of-friends" that can be revoked at any point?