BP-Rec. 1: Use a unique naming convention for concept/class and relations

[ghost]

BP-Rec. 1: Use a unique naming convention for concept/class and relations
Description Concept/class and relations composing semantic artefact are associated with a human readable label which is a character chain. This chain can become complex and include several associated words (e.g. Hyperplastic and giant kidney ). There are multiple conventions for naming semantic artefact elements such as CamelCase or the conventions proposed by OBO Foundry and Industry Ontology Foundry. Unfortunately these existing conventions/recommendations are not harmonized which leads to the need for a search engine or an automated mapping service to comply with the different conventions/recommendations. This hampers both searching capabilities and automated mappings. This recommendation for Best Practice emphasizes the need to define a common unique naming convention by the community of practitioners.
Existing recommendations : ● OBO Foundry - Principle 12 (Schober et al., 2009) [1] ● Industry Ontology Foundry - requirement 11 [2]
Stakeholder : Practitioner and Community

[EamdouniGIT]

Ideally, repositories should adopt a naming mechanism to attribute a unique global prefix for a semantic artifact. A centralized prefix service is needed for reserving namespaces and assigning globally unique and persistent URIs. The AgroPortal repository assigns to ontologies an acronym identifier which is unique but not universal (e.g., ENVO, PPO, AGROVOC). This acronym (stored in omv:acronym) is usually chosen by the author or AgroPortal's administration with respect to the best practices in certain communities (OBO Foundry for example). @jonquet any other ideas?

[graybeal]

I believe the previous comment is about the artifact rather than the concept/class labels, which I think this is addressing. Perhaps it is suitable for BP-Rec. 2?

I think I would frame the goal here as "emphasizes the value of having a recommended labeling convention." These are subtle changes throughout, justified below.

• Many people think the 'name' is the identifier; this is not about the identifier.)
• Often the labels are chosen to meet specific needs for a vocabulary and its terms, that is the more immediate need for those users
• There are so many reasons to organize names differently, choosing just one will seem rather arbitrary
• There are multiple languages involved (there can be multiple labels for each term, one per language)
• So a common convention (one that everyone follows) seems quite unlikely
• Uniqueness seems out of place here. Of course it will be unique (except that it will also be common). Suggest dropping it, or clarifying it.