You are reading content from Scuttlebutt
@Anders %M9y2zjKV4CY9nEXsQEy2edwttzKa0aKqyAllpWmEI8E=.sha256
Re: %mamZPa6Zi

The main question I have, that I’m not clear on, is how partial replication changes validation of the messages. One can still validate the signatures of the messages but not the other properties. For example how does one know a feed it forked?

When doing partial replication using indexes it is true that you can't know for sure whether the author left out messages. We had a design that could help with that, but its complex. What you know for sure is that the author wrote the index and the indexed messages.

I see indexes as mainly something we use for classic main feeds. Going forward we now have: a way to introduce more feed formats and to split data into multiple feeds for different use cases, like contacts and specific applications (git, etc.). And these feeds will have the normal linked properties, might even have more properties when using something like bamboo.

As for the forking aspect that is in general a very hard problem to solve. I don't think the addition of metafeeds changes that very much. It was a problem for classic feeds and is still a problem. What I think might help this is when we get fusion identities implemented. Fusion identities means your never really have to share keys. Leaving the forked feed mainly a problem of backup. And in that case I think a proper: "save my feed & latest sequence(s)" + restore from that and only allow posting once you have all that data should be safe.

Join Scuttlebutt now