I have a new random question: Has anyone seen attempts to take a functional language and mix it with a graph database / query language? Today I've been wondering if you can have a standard "select" query from a graph DB spit out the resultant subtree(s) as an ADT (record) that a functional language can traverse/manipulate as normal in order to make an update. You could take the Elm Architecture and just swap out the Model for a graph DB, and have the update function trigger a query upon it that feeds data into the functional language.

The definition of a functional language is rather vague, so it isn't easy to answer your question. My Beads project uses a graph database (Neo4J is the closest shipping product to what i have built), and is what i call hyperfunctional, as it uses many tiny functions that are orchestrated by the runtime. I am targeting general purpose interactive graphics applications for the web, mobile and desktop. Trying to make what some people call an executable specification; where the spec is the program.

this is kind of how graphql and datomic work

@Garth Goldwater I've been intending to take a deeper look into the datomic API 🤔

Possibly relevant. Used more for code analysis but I don't see why it couldn't be taken in reverse direction. https://github.com/ShiftLeftSecurity/codepropertygraph