From graphs to tables
|SMWCon Fall 2016
|From graphs to tables
|The case for storing semantic data in tables instead of, or in addition to, graphs - for SMW/Cargo, Wikidata, and beyond
|Cargo, Wikidata, templates
We are used to thinking about "Semantic" data in terms of graphs - nodes and arrows, stored as triples in RDF and queried in SPARQL. But there's no rule that says that data on the so-called Semantic Web, or part of the Linked Open Data cloud, etc. has to be represented that way. The Cargo extension for MediaWiki is an alternative to Semantic MediaWiki that stores its data in true relational database tables rather than as RDF triples. (To be sure, in most cases SMW also stores its data in relational DB tables, but they're "representing" a triplestore.) One of the main advantages given for triplestores is that they can hold a flexible data structure, but the Cargo extension provides flexibility in its own way through dynamically-generated tables. This talk will cover the Cargo approach, as well as general benefits of using relational databases and SQL instead of, or in addition to, triplestores and SPARQL; and will also offer some possibilities for the future of Wikidata and similar efforts.