Currently DWT runs on a single machine. Federated queries such as "What expressions on any of my friends' computers involve wedding, invitation and Smith?" would be useful.
Some powerful pattern-matching graph query languages (Sparql, Gremlin, Datalog ..) already exist. Implementing DWT's (easier) query language using those would make it more expressive and faster.
Among open-source graph databases, JanusGraph is uniquely scalable, because it offers vertex-centric indices. It uses Gremlin, a meta-language, so if we decided to switch to something else like Neo4j, it would take minimal work.
Among Haskell web toolkits, Servant is magically beautiful, but I haven't thoroughly studied the alternatives.
Currently DWT runs on a single machine. Federated queries such as "What expressions on any of my friends' computers involve
wedding
,invitation
andSmith
?" would be useful.Some powerful pattern-matching graph query languages (Sparql, Gremlin, Datalog ..) already exist. Implementing DWT's (easier) query language using those would make it more expressive and faster.
Among open-source graph databases, JanusGraph is uniquely scalable, because it offers vertex-centric indices. It uses Gremlin, a meta-language, so if we decided to switch to something else like Neo4j, it would take minimal work.
Among Haskell web toolkits, Servant is magically beautiful, but I haven't thoroughly studied the alternatives.