3 ms·
I've had to dive into this recently as part of my work (I'm in academia). While the original vision seems to have failed from a lack of incentives, the idea be
by simongray 6y ago
I've had to dive into this recently as part of my work (I'm in academia).
While the original vision seems to have failed from a lack of incentives, the idea behind RDF as way to model connected data and its many serialisation formats is actually quite helpful. It seems to have found a niche in academia when it comes to sharing and integrating large graph datasets. I just consider RDF and most of the other semantic web standards as a framework for modelling and serialising graph data. It's pretty good at that.
- tasogare 6y agoNo, it’s not even good for that. The fact you need yo reify properties (relationships) to annotate them is a big failure, which is not present in more general graph models.
- Karrot_Kream 6y agoI think this is a matter of opinion. I'm personally a fan of reified properties, but this is the source of differences between property graph databases and RDF triple stores.
- simongray 6y agoWorks fine for my use cases (e.g. representing multiple WordNets as labeled property graphs in Neo4j) and I quite like the simplicity of data modelling based on triplets of data literals. What do you suggest as an alternative to RDF?
- emmanueloga_ 6y agoNot true. See "reification is a red herring" [1]. Also, there's now a more natural way to add properties to relationships if really needed, it is call RDF* [2], and it is already implemented of a lot of triplestores, including RDF4J, Jena, StarDog and more. 1: http://www.snee.com/bobdc.blog/2018/04/reification-is-a-red-herring.html http://www.snee.com/bobdc.blog/2018/04/reification-is-a-red-... 2: https://blog.liu.se/olafhartig/2019/01/10/position-statement-rdf-star-and-sparql-star/ https://blog.liu.se/olafhartig/2019/01/10/position-statement...