Hello everyone,
Daniel Kuppitz did some really cool work getting SPARQL to compile to Gremlin's
instruction set.
https://github.com/dkuppitz/sparql-gremlin
http://www.datastax.com/dev/blog/the-benefits-of-the-gremlin-graph-traversal-machine#section2
Right now, this work is in Kuppitz' personal repo. However, I was thinking that
it would be good to boost up this project such that we may someday merge it
into Apache TinkerPop.
Why?
1. We have 3 reference graph system implementations in TinkerPop.
- TinkerGraph (POJO -- OLTP/OLAP)
- Neo4j (OLTP)
- Hadoop (OLAP)
2. We have 2 reference graph language implementations in TinkerPop.
- Gremlin-Java8
- Gremlin-Groovy
Why not have 3 language reference implementations AND have SPARQL-Gremlin be
that 3rd language because:
1. It would demonstrate a language that is not embedded in a
host-language (e.g. Java8, Groovy, Scala, etc.)
2. It would give TinkerPop an automatic RDF story without us having to
do the whole Sesame/Jena song-and-dance.
Talking with Kuppitz, he likes the idea but is not too interested in
deep-diving into SPARQL and spending too much of his time on this body of work.
Perhaps someone out there in TinkerLand is an expert in SPARQL, understands the
visitor-pattern used in the SPARQL-Gremlin compiler (uses Apache Jena's ARQ
parser), and is willing to fully flesh out all the features of SPARQL into
Gremlin.
Hope someone bites.
Take care,
Marko.
http://markorodriguez.com
