I’m really interested in this!
Why scuttlebutt vs swarm.js?
How would you scale vertically and horizontally? Any experience with plugging it into a message queue or pubsub model?
Most CRDTs aren’t document-based, for example instead of editing a field in a JSON, I’d edit a string of text as position X. This looks like a fine first step to moving towards a document-based structure. Any thoughts on how far to take it? Possible integration with GraphQL or other document schemas?
If i have ephemeral domain states (eg socket presence) and persisted domain states (eg shared document) do you recommend to feed everything into scuttlebutt, or would you send persisted stuff right to the DB & just use scuttlebutt for the ephemeral stuff?
This is what the web needs, thanks for pushing us there!