Semantic artefacts are living digital entities undergoing changes and revisions to cope with semantic drift and for improving/extending the scope or granularity. Provenance information describing all these changes during the semantic artefact lifecycle should be provided to potential external users. This information can be thus used to evaluate the semantic artefact and understand the release cycle.
Provenance should be documented at an appropriate level of granularity to enable reuse of semantic artefacts and its constituting elements (class/term and relation). Provenance should be presented to the human user but also should be expressed in a machine-readable way. All appropriate sources should be referred to (both source pid (data object) and creator pid) and the provenance should provide dates and lifecycle events. The provenance information should be described using an appropriate standard model such as PROV[53]. PROV-based machine readable description could be then used to provide means to automatically update any resource using the semantic artefact. The provenance information should contain all the necessary elements to build representations to the users such as changelogs and describe the backward interoperability.
Assessing if a semantic resource follows community standard is hard because one needs to look deeper inside an ontology and evaluate how it is built in relation to state-of-the-art ontologies and practices. However, one can simply recognize the use of a standard language (e.g., OWL or SKOS) or the inclusion in certain curated/guided ontology frameworks like the OBO Foundry initiative for biological and biomedical ontologies.
We are inclined, in an open science vision, and because it is a community practice, to recognize a higher level of FAIRness for semantic artefacts that are fully open and accessible (even if a privately restricted ontology can be very FAIR too).
AgroPortal recognizes the inclusion of ontologies in certain groups corresponding to use cases or communities. For example, AgroPortal has a group of ontologies for all the semantic resources developed by INRAE scientists. We also have a group to capture the inclusion in the OBO Foundry or Crop Ontology project. When the inclusion in a group recognize the adoption of certain domain-relevant community standards or practices.
Concretely, R1.3 is partially assessed by I1 in relation to format/language used. In complement, we can assess R1.3 by evaluating if an ontology is included in a group of ontologies respecting certain community practices (in AgroPortal this will include: OBOF, CROP, INRAE). We can also recognize openness via the metadata bpm:viewingRestriction.
@EamdouniGIT Profiles are your friend here.. multiple profiles with specialisation heirarchies allow common aspects to be defined and simple tests implemented (eg. SHACL). Trying to create a "super standard" covering all possibilities is probable impossible.
Description
Semantic artefacts are living digital entities undergoing changes and revisions to cope with semantic drift and for improving/extending the scope or granularity. Provenance information describing all these changes during the semantic artefact lifecycle should be provided to potential external users. This information can be thus used to evaluate the semantic artefact and understand the release cycle.
Provenance should be documented at an appropriate level of granularity to enable reuse of semantic artefacts and its constituting elements (class/term and relation). Provenance should be presented to the human user but also should be expressed in a machine-readable way. All appropriate sources should be referred to (both source pid (data object) and creator pid) and the provenance should provide dates and lifecycle events. The provenance information should be described using an appropriate standard model such as PROV[53]. PROV-based machine readable description could be then used to provide means to automatically update any resource using the semantic artefact. The provenance information should contain all the necessary elements to build representations to the users such as changelogs and describe the backward interoperability.
Existing recommendations:
- MIRO (Matentzoglu et al., 2018)
- OBO foundry - Principle 4[54]
- OBO Foundry - Principle 8[55]
- PROV
Stakeholder: Practitioner
@jonquet
Assessing if a semantic resource follows community standard is hard because one needs to look deeper inside an ontology and evaluate how it is built in relation to state-of-the-art ontologies and practices. However, one can simply recognize the use of a standard language (e.g., OWL or SKOS) or the inclusion in certain curated/guided ontology frameworks like the OBO Foundry initiative for biological and biomedical ontologies. We are inclined, in an open science vision, and because it is a community practice, to recognize a higher level of FAIRness for semantic artefacts that are fully open and accessible (even if a privately restricted ontology can be very FAIR too).
AgroPortal recognizes the inclusion of ontologies in certain groups corresponding to use cases or communities. For example, AgroPortal has a group of ontologies for all the semantic resources developed by INRAE scientists. We also have a group to capture the inclusion in the OBO Foundry or Crop Ontology project. When the inclusion in a group recognize the adoption of certain domain-relevant community standards or practices.
Concretely, R1.3 is partially assessed by I1 in relation to format/language used. In complement, we can assess R1.3 by evaluating if an ontology is included in a group of ontologies respecting certain community practices (in AgroPortal this will include: OBOF, CROP, INRAE). We can also recognize openness via the metadata bpm:viewingRestriction.
@EamdouniGIT Profiles are your friend here.. multiple profiles with specialisation heirarchies allow common aspects to be defined and simple tests implemented (eg. SHACL). Trying to create a "super standard" covering all possibilities is probable impossible.
+1 in general - see #22