OntoPortal Content Identification (URIs, DOIs, PURLs, ...)

[syphax-bouazzouni]

Definition of Dereferencing

In the context of web semantics, dereferencing refers to the process of resolving and obtaining the actual data associated with a Uniform Resource Identifier (URI). In simpler terms, it involves following a link represented by a URI/IRI to retrieve information or resources linked to that identifier.

In the Semantic Web, URIs are used to uniquely identify resources, and dereferencing allows systems to access and retrieve data related to those resources. When a URI/IRI is dereferenced, it typically leads to the retrieval of RDF (Resource Description Framework) data or other structured information that describes the resource in a machine-readable format. This enables systems to understand and process the meaning of the linked data, facilitating the exchange and integration of information on the web.

Definition of Content negotiation

Content negotiation in the context of the Semantic Web refers to the mechanism by which two communicating parties, such as a client and a server, agree on the most suitable representation of a resource during the process of dereferencing a URI or IRI. This negotiation aims to ensure effective communication between different systems that may prefer different data formats or languages.

In other words, when a client requests a resource by dereferencing a URI, it indicates its preferences for the format or language of the response data through HTTP headers or other negotiation mechanisms. The server, in turn, examines these preferences and selects the most appropriate representation of the resource based on what is available.
different formats can be agreed upon between a client and server when accessing linked data. Common formats include:

• RDF/XML: XML-based representation of Resource Description Framework data.
• Turtle: Human-readable serialization format for RDF.
• JSON-LD: JSON-based format for linked data.
• N-Triples and N-Quads: Text formats for expressing RDF triples and quads.
• HTML: Markup language for web pages, also used to embed RDF data.
• RDFa: Embedding RDF data in HTML or XML using attributes.
• SPARQL Query Results XML and JSON: Formats for representing SPARQL query results.

Problematic

In Agroportal, of the 176 ontologies, we have

• 17 (9%) ontologies with no URIs.
• 62 (35%) ontologies with URIs that do not work (return error status).
• 84 (48%) resolves but are not content negotiable, supporting only one format, 51 been html, 27 plain text, one xml, and one rdf+xml.
• 4 (2%) resolves and supports 2 different formats (mostly html, json, plain textor rdf+xml formats).
• 3 (2%) resolves and supports 3 different formats (mostly html, json, ld+json, turtle or rdf+xml formats).
• And finally 6 (4%) resolves at least 4 different supports.

In conclusion in Agroportal, only 13 resources (8%) are fully resolvable and content negotiable

Goals

1. Human-Readable URIs: Our primary goal is to assign human-readable URIs to all resources hosted in an Ontoportal instance. For example, the ontology may have a URI like https://agroportal.lirmm.fr/ontologies/AGROVOC, and a concept/class could be identified by https://agroportal.lirmm.fr/ontologies/AGROVOC/c_330834. We commit to maintaining and ensuring content negotiation for these URIs, offering accessibility in various formats, including XML, CSV, Turtle, HTML, and JSON.

2. URI Authority Management: For resources already utilizing URIs (domain names) that currently do not resolve or need redirection to an Ontoportal instance, our goal is to provide utilities and guidance. We aim to assist them in transferring authority over their URIs to us. Additionally, we are exploring the possibility of automating this process if technically feasible (pending further investigation).

to be continued

Roadmap

• Agroportal Internal URI derefrencement agroportal/project-management#476

to be continued

[galviset]

I think that feature is also important for other reasons :

• I did come across users that would need to get dereferenced URIs but cannot provide their own or already have one but wish to switch platforms
• there is a whole debate about the human-readability of URIs. However I think that providing easy, logically formed URIs makes it easier for both the average user and any kind of developer that needs to work with them
• this provides a better distribution of ontologies overall and this is part of the philosophy of Ontoportal