SSB's root is just "hash chain + signatures", and is relatively trivial to implement. Most of the actual protocol is flawed in several ways, and there's tough issues they are still trying to solve. I'd say something based off of IPFS for bulk storage (SSB suffers from bloat, and there's multiple proposals to move actual data off the append only log and into "blobs", which are already WAY better implemented in IPFS, IMHO), with some form of "here's the bits that make up my profile" thing, possibly similar to SSB, but with a cleaner implementation. A few things that should be high priority, IMHO:
- Allow for profile "compaction"/data removal (SSB says you should never delete anything, which is frankly untenable, IMO)
- Allow for multiple devices/apps to manipulate your profile concurrently (Complex, but SSB has been criticized multiple times for vigorously arguing against this capability)
- Use a data protocol that is well formed and structured (SSB uses "node.js formatted" json serialization, and then verifies signature after deserializing and reserializing, meaning you must match node.js's serialization perfectly to interact with the rest of the ecosystem)
Re:Reinvent the Social Web
by crawfordcomeaux on Thursday October 18, 2018 @10:25PM
(#18253868)
I'm in. Gonna start researching the stack while waiting for my firstborn to arrive. What's next for you?
Hope that someone builds it for me ;). I can't really do anything with that concept, because life is very full for me currently and is likely to remain so for the next year. It's just something I've been scribbling down on paper and prodding at occasionally. I suspect it's a semi-natural idea that'll occur to a lot of people when they start looking at Dat, IPFS, SSB, SOLID, ActivityPub etc. and trying to find where commonalities may exist.
It seems like an impossibly vast and intimidating project. I think the main takeaway from this ramble is that I think the new tools emerging to re-decentralize the web are really, really cool :).