Node started before JS modules where standardized, also they didn't adopt promises right away and favored callback instead.

I find things like pull-streams (congrats to your baptism, btw) very elegant, but to write understandable code I prefer async iterables with the built in support of async-generators and for-await loops.

I struggle a bit with the fact that you referrer to the node stack as "the JS stack" as deno is much closer to standard JS than nodejs, the scuttlesaurus code can run in the browser without browserify or similar. But apart from such naming issues, I think it would be great if there were more implementations of scuttleverse portal as I think they can make scuttlebutt useful for more people and communities.

I've started a repo on GitHub that might evolve to a specification of scuttleverse portals: https://github.com/retog/scuttleverse-portal

Things like RDF and SPARQL aren't necessary for a scuttleverse portal, the ability to sync with another (ws) peers and to export the key should however be common to every portal.