What I like
- The prominent use of Named Graphs.
If you import triples into the store you need to supply a graph id. Automatically another graph is created, a metadataGraph containing metadata statements about the first graph.
In the graph with id "http://www.example.com/graph1" indicated in green you have your statements.
In the associated metadatagraph "http://openanzo.org/metadataGraphs(http%3A%2F%2Fwww.example.com%2Fgraph1)" you will find statements about the "http://www.example.com/graph1" graph and the metadatagraph itself. It is obvious that this mechanism can be used for all kinds of purposes such as indicating provenance, managing access rights, ... Nice.
- the CLI.
The Command Line Interface offers an easy, intuitive way of interacting with the graphs and the store.
- the GLITTER engine for executing SPARQL queries and the extensions to SPARQL it supports.
This makes some type of queries much, much easier to do. Bear in mind however that these are not standard SPARQL facilities, although I do expect some of these will appear in the next version of SPARQL, Lee Feigenbaum being involved.
- the replication service
Open Anzo's replication service offers 2-way synchronization between Anzo clients and an Anzo server.
What I don't like
- there is no REST API for the moment; a must have for us.
- no built-in reasoning yet.
- bulk loading seems to be slow.
- there seems to be very, very little activity in the openanzo google group, which is a sign of not being used much.
We have been there before, being a lonely user of software; we will not make this mistake again.
Living in the XML and RDF world
Comments