Protocol ontology, taxonomy, and a catalog

I was looking for a protocol ontology, taxonomy or a catalog and couldn’t find any.

Then I hoped to at least find “Protocol” type/class somewhere. None in[1]. Searching in well-documented ontologies, I was able to find a Protocol class only in some obscure Informatics Research Artifacts Ontology.

Wikidata, otherwise quite reliable, when it comes to protocol classes is a mess. There are different protocol classes defined as etiquette of diplomacy, operating procedure, transcript of an observed sequence of events, predefined method, treaty and so on. Things look a bit better for “communication protocol”. There are over 8000 items (← the link works courtesy of the powerful SPARQL protocol) classified either directly as communication protocol or as a subclass of it. Some are probably wrongly typed but it is already a good dataset to improve and extend.

In DBpedia ontology there is no protocol class at all but searching dbr:Protocol as connected hypernym gives some results.

These (see subject) might be a valuable, tangible things to work on. Since it’s not a PIG or a PILL, if others also see value in it, it might be included in the next year’s Summer of Protocols.

Some initial thoughts regarding a technology stack:

A well-established approach is to use W3C standards.

The ontology can be expressed in RDFS if a lightweight one or in OWL if more axioms are needed.

The validation chicha can be created in SHACL, and the taxonomy in SKOS. The tags in this index are seeds of such a taxonomy.

The catalogue itself can be decentralized, inviting people documenting protocols to publish a structured description standardized by the agreed-upon ontology and taxonomy, similar to what Google does for local businesses, books and suchlike, and embed it as JSON-LD or RDFa in their publications (protocols specs).

Aggregating these descriptions can be done with Walder or a similar open-source tool.

What do you think?

  1. But it seems they are open to include it. ↩︎


I have an idea, but let me think about it first.