[taurknaut]

> rather than abusing existing relational databases for graph purposes.

In my experience, the vast majority of graphs can be embedded in relational databases just fine and most people don't want general graph querying. People just don't like optimizing queries (or equivalently the schema to enable such queries).

I personally have never seen a pitch for graph databases that makes them seem attractive for more than data exploration on your local machine.

[voodooEntity]

Well im working since multiple years on a private lets call it "research" project which deeply relies on growing/deep structured graphs.

I don't think that GraphDBs should a default choice, but there are cases in which they just perform better.

Could i write my research project with a relational DB? Yes - i tried - and it sucked xD

[hobs]

Yep, something like adjacency lists have solved a lot of dumb recursive queries/"graph queries" for me, usually because the graph doesn't change much and is just a few nodes deep.