Naga actually just wraps a store. The idea was that it could wrap any graph at all, starting with Datomic and SPARQL. But then I was asked to do something myself (before I got to the SPARQL implementation) and that’s how Asami got started

@quoll: One other thing; is it necessary for asami/naga to provide a registry and connection strings etc? Is there an interface where I can just provide a db as a value? Also that way if I wanted mutability or a registry; I could provide my own atoms, rather than you encapsulating them behind an identifier etc?

You’re talking about the Graph interface (well… protocol)

Connection/Database is just a wrapper around this

Connections associate a URI with the Graph instance

and then provide a history as a series of databases. Databases are VERY thin wrappers around a graph

Until a couple of months ago, there was only the Graph. But everyone loves Datomic, so I wrapped it to look like Datomic 🙂

Naga wrapped the graph, but only does so in order to manage transactions. The functions attached to the wrapper were useful for Datomic as well, so I kept them at that level

Asami Graph instances can be inserted into, deleted from, and queried. And until recently, that was everything we did with them. I made it messy by giving it a Connection. But you can still pull the graph out. And if you want to build a connection around a graph, then you can do that too

It just occurred to me… This is all VERY Asami specific. Maybe it’s better to talk about this in #asami

:thumbsup: sure didn’t realise there was an #asami 🙂 thanks

🙇 forgot to say thanks a million for all the answers @quoll