RDF is a data model for Linked Data specified by the W3C.
Data is modeled as a graph of nodes with labeled directed edges.
RDF is quite widely used, however it is not very popular.
Criticisms include (unnecessary) complexity (see also Semantic Web criticism).
Here some reasons why it may still be worthwhile to consider RDF as data model.
Ontologies or (vocabularies) describe a formalization of certain real-world concepts.
Ontologies are always subjective to a certain view of the world. RDF explicitly allows this subjectivity by allowing the creation and usage of custom ontologies. To ensure interoperability it is necessary to reuse common ontologies.
Many people and projects have been working on many different ontologies for specific purposes. Being able to reuse this work is immensely valuable.
Two underlying foundational principles make RDF extremly well suuited for distributed and decentralized systems:
- Unique name assumption: Everything has a unique name and is referenced by that name.
- Open-world assumption: Not everything is known to individual nodes. The truth value of a statement may be true but not be known to be true.
The foundations of RDF are well understood and well described (see Simple and Efficient Minimal RDFS (2009)).
RDF 1.1 Concepts and Abstract Syntax
The central RDF specification.
Simple and Efficient Minimal RDFS (2009)
A primer on the Semantic Web and Linked Data (2019)
An introduction to the Semantic Web and Linked Data.
The "problem child" of RDF.
A selection of interesting ontologies.