Semantic Web Layer Cake Core Technologies/Frameworks and Related: A Trends Comparison – (Post Under Construction)

Spread The Knowledge...

Fig. 1 Updated Semantic Web Layer-Cake According To [1]

We introduce this technical report with the main components of the updated Semantic Web layer-cake technologies/frameworks according to [1] and also include related graph-based technologies such as query languages and knowledge graphs:

  • RDF Document Serialization Formats: RDF/XML, RDFa, RDF N-Triples, RDF Turtle, RDF Trig, JSON-LD, N-Quads, N3-Notation3, TRIX, Hex-Tuples, HDT, RDF*
  • Ontologies/Vocabularies: RDF Schema (RDFS), Web Ontology Language (OWL), Simple Knowledge Organization System (SKOS), Schema.org, Dublin Core, Friend-Of-A-Friend (FOAF), Semantically-Interlinked Online Communities (SIOC), Gene Ontology (GO), Systematized Nomenclature of Medicine—Clinical Terms (SNOMED CT), Coronavirus Infectious Disease Ontology
  • Rule-Based Languages/Standards: Semantic Web Rule Language (SWRL), SPARQL Inferencing Notation (SPIN), RDB to RDF Mapping Language (R2RML), Rule Interchange Format (RIF), Shapes Constraint Language (SHACL)
  • Query Languages:
    • Semantic Web Query Languages: SPARQL Protocol and RDF Query Language (SPARQL), SPARQL*/RDF* (Property Graphs), SPASQL
    • Graph Query Languages: GraphQL, Cypher/Open Cypher, Gremlin, Graph Query Language (GQL)
  • Semantic Web Ontologies vs Knowledge Graphs
  • Knowledge Graphs:
    • Open Knowledge Graphs: DBpedia, Wikidata, YAGO, Freebase, BabelNet, KBpedia, CaLiGraph, NELL, Opencyc
    • Commercial Knowledge Graphs: Google Knowledge Graph, Google Scholar, Bing Knowledge Graph,  LinkedIn Knowledge Graph, Microsoft Academic Graph, Facebook Knowledge Graph, DiffBot Knowledge Graph, Cyc Knowledge Base, Amazon Product Knowledge Graph, Walmart Product Knowledge Graph, E-Bay Knowledge Graph

Fig. 2 Semantic Web Layer-Cake Core Technologies/Frameworks and Related Technologies

RDF Serialization Formats Section Ontologies/Vocabularies Section Rule Based Languages Standards Section Query Languages Section Knowledge Graphs Section Open Knowledge Graphs Sub-Section Commercial Knowledge Graphs Sub-Section Semantic Web Query Languages Sub-Section Graph Query Languages Sub-Section

RDF Serialization Formats

In this section of this technical report, we include 12 of the most popular/prominent RDF serialization formats, starting with the original/oldest RDF/XML RDF serialization format, the scattered in the “XHTML document” RDFa serialization format, the currently most popular JSON-LD serialization format, the human-readable/compact RDF Turtle serialization format, the performant Hex-Tuples serialization format, the compact/compressed/binary HDT serialization format, the RDF*/Turtle* serialization format that permits the definition/bridge to the Labelled Property Graphs (LPG) model among others:

  • RDF/XML: RDF/XML 1.1 a W3C Recommendation on the 25 February 2014 is an XML syntax for expressing RDF graphs as XML documents. It was the very first syntax for expressing RDF Graphs and it is more focused on expressing machine-readable RDF graphs via verbose representations.
  • RDFa: RDFa 1.1 is a W3C Recommendation on the 17th of March of 2015 used for the purpose of adding attribute-level extensions to HTML, XHTML, and XML-based document types for embedding RDF data.
  • RDF N-Triples: RDF N-Triples 1.1 a W3C Recommendation also on the 25 February 2014  being a line-based, plain text serialization format for RDF graphs, and a subset of the Turtle format.
  • RDF Turtle: RDF 1.1 Turtle Terse RDF Triple Language is a W3C Recommendation 25 February 2014 whose syntax is similar to that of SPARQL written in a compact and natural text form, with abbreviations for common usage patterns and datatypes. It is a very popular RDF serialization format being human-readable and a superset of the line-based N-Triples format.
  • RDF TriG: RDF 1.1 TriG a W3C Recommendation also on 25 February 2014, written in a compact and natural text form, with abbreviations for common usage patterns and datatypes being an extension of the Turtle serialization format.
  • JSON-LD: a lightweight Linked Data format which currently is a W3C Proposed Recommendation 07 May 2020, human-readable/writing based upon the JSON format and providing a data format ideal for programming environments, REST Web services, and document-based No-SQL databases.
  • N-Quads: a W3C Recommendation also on 25 February 2014 which is line-based, plain text format for encoding RDF data in Subject-Predicate-Object sentences also containing a 4th optional component indicating the graph in a dataset the triple belongs otherwise the triple belongs to the default graph
  • N3-Notation3: N3 a W3C submission that was designed with human-readability consumption and developed by Tim Berners-Lee and others from the Semantic Web community. N3 has several features that go beyond the serialization of RDF models from which RDF-Turtle is a simplified RDF-only subset of N3.
  • TRIX: Triples in XML is a practically deprecated XML-based serialization format for RDF Named Graphs and RDF Datasets which offers a compact and readable alternative to the XML-based RDF/XML syntax.
  • Hex-Tuples: HexTuples is a performant RDF serialization format/simple data model for dealing with linked data. A HexTuple consists of six fields: subject, predicate, value, datatype, language, and graph. It benefits from highly optimized JSON parsers in browsers.
  • HDT: HDT (Header, Dictionary, Triples) is an open-source/in progress of standardization (W3C HDT Member Submission) compact data structure and binary serialization format for RDF that compresses RDF datasets saving space while maintaining search and browse operations without prior decompression. Most parts of the data can be kept in main memory, which makes it an ideal format for storing and sharing RDF datasets on the Web. The size of the files is much smaller which means less bandwidth consumption for both producers/consumers. HDT is read-only which enables the use of fast/performant multi-threading programming. 
  • RDF*:  RDF star or extended RDF covers the lack of a convenient way to annotate RDF triples and to query annotations as done natively in other graph data models such as edge properties in the “Property Graph” model. Whereas some previously proposed solutions relied on making use of reification annotating RDF statements with certainty scores, weights, temporal restrictions, and provenance information, these suffered from being too verbose and redundant. In [2] is proposed RDF*/Turtle*/SPARQL* as an alternative approach that is based upon the nesting of RDF triples and of query patterns allowing a compact representation of RDF data and SPARQL queries, which is backward-compatible with the RDF/SPARQL standards, backed by a solid formal theoretical model [3,4,5], permitting the interoperability/conversion between the triple-based RDF model and the Property Graphs models, and also permitting more specific annotation-related extensions of RDF and SPARQL such as temporal or probabilistic extensions. RDF* extends RDF by allowing arbitrarily deeply nested triples given the assumption that triples representing/annotating metadata about another triple may directly contain the annotated triple as its subject or object. In other words, RDF* permits the non-verbose representation of data and the corresponding metadata, in the form of nested triples.

In the appendix, we include/show examples of each of these 12 RDF serialization formats.

Here we include Google Trends for the most prominent serialization formats in the latest 5 years.


It can be seen that the JSON-LD serialization format has been substantially more popular (in the latest 5 years) than the rest of serialization formats. The second most popular serialization format is RDF Turtle but far apart from JSON-LD. Most probably is that JSON-LD has been more popular among developers given that is both readible for humans and easily consumed by machines. RDF Turtle is quite popular in the Semantic Web Academic arena

Semantic Web Ontologies/Vocabularies

Ontologies/Vocabularies: RDF Schema (RDFS), Web Ontology Language (OWL), Simple Knowledge Organization System (SKOS), Schema.org, Dublin Core, Friend-Of-A-Friend (FOAF), Semantically-Interlinked Online Communities (SIOC), Gene Ontology (GO), Systematized Nomenclature of Medicine—Clinical Terms (SNOMED CT), Coronavirus Infectious Disease Ontology

Explain RDF Schema, Web Ontology Language, SKOS, Schema.org and identify in Semantic Layer cake

  • RDF Schema (RDFS): RDF Schema stands for Resource Description Framework Schema. The current and latest version is RDF Schema 1.1 which is a W3C Recommendation on 25 February 2014. RDF Schema is normally being used to define vocabularies and taxonomies allowing the following primitives, class/subclass definition, property/sub-property definition, and utility properties. It is also used as a basis in the Web Ontology Language (OWL)
  • Web Ontology Language (OWL): OWL 2 ontologies enable the definition of classes, properties, individuals, and data values. It is a W3C Recommendation on 11 December 2012. With OWL it is possible to define class/property hierarchies, complex class relationships such as class disjointness, property restrictions and property cardinality restrictions, property chaining and so on (@@@add more)
  • Simple Knowledge Organisation System (SKOS): SKOS is a W3C Recommendation on 18 August 2009. It is used “per-se” or in combination with OWL in order to define common data models for knowledge organization systems such as thesauri, classification schemes, subject heading systems, and taxonomies.
  • Schema.org: although not a W3C recommendation it is a defacto standard. It is a collaborative effort with a mission to create, maintain, and promote schemas for structured data on the Internet, on web pages, email messages, organizations, places, persons, products, community extensions, and so on. Over 10 million Web sites use Schema.org to markup their content. Schemas are classified hierarchically in a set of types/properties adding a total of 829 types, 1351 properties, and 339 enumeration values. The latest release is version 8.0 on 2020-05-01

Semantic Web Rule-based Languages/Standards

Rule-Based Languages/Standards: Semantic Web Rule Language (SWRL), SPARQL Inferencing Notation (SPIN), RDB to RDF Mapping Language (R2RML), Rule Interchange Format (RIF), Shapes Constraint Language (SHACL)

Explain the different Semantic Web Rule-based query languages and how do they fit in the Semantic Web Layer Cake It can be seen that Knowledge Graphs are more popular in China and in United States Ontologies and Knowledge Graphs are similarly distributed. Therefore it is not by surprise that a lot of research papers related to Knowledge Graphs are submitted by Chinese researchers.

Query Languages: Semantic Web Query Languages

Semantic Web Query Languages: SPARQL Protocol and RDF Query Language (SPARQL), SPARQL*/RDF* (Property Graphs), SPASQL

Query Languages: Graph-based Query Languages

Graph Query Languages: GraphQL, Cypher/Open Cypher, Apache TinkerPop™ Gremlin Graph Traversal Language, Graph Query Language (GQL)

Explain the different Semantic Web query languages SPARQL and SPASQL and graph-based query languages Gremlin, Cypher and Graph-QL It can be seen that Knowledge Graphs are more popular in China and in United States Ontologies and Knowledge Graphs are similarly distributed. Therefore it is not by surprise that a lot of research papers related to Knowledge Graphs are submitted by Chinese researchers.

Ontologies vs Knowledge Graphs

Explain and define what is an Ontology and what represents and define Knowledge Graphs It can be seen that Knowledge Graphs are more popular in China and in United States Ontologies and Knowledge Graphs are similarly distributed. Therefore it is not by surprise that a lot of research papers related to Knowledge Graphs are submitted by Chinese researchers.

Knowledge Graphs: Open Knowledge Graphs

Open Knowledge Graphs: DBpedia, Wikidata, YAGO, Freebase, BabelNet, KBpedia, CaLiGraph, NELL, Opencyc

  •  
Knowledge Graphs: Commercial Knowledge Graphs

Google Knowledge Graph, Google Scholar, Bing Knowledge Graph,  LinkedIn Knowledge Graph, Microsoft Academic Graph, Facebook Knowledge Graph, DiffBot Knowledge Graph, Cyc Knowledge Base, Amazon Product Knowledge Graph, Walmart Product Knowledge Graph, E-Bay Knowledge Graph

References

[1] Kingsley Uyi Idehen, “Semantic Web Layer Cake Tweak, Explained”, Medium 2017 Accessed at: https://medium.com/openlink-software-blog/semantic-web-layer-cake-tweak-explained-6ba5c6ac3fab

Print Friendly, PDF & Email

Spread The Knowledge...