What does mod:SemanticArtefact "describe/annotate"? (and other questions)

[dgarijo]

I am trying to use MOD for summarizing the findings of an analysis. I have the following questions, which I have not been able to answer after looking at the figures like this one and the documentation.

• I understand what a mod:SemanticArtefact is, but not that well what is the "mod:annotates/describes" property (linking it to a Dataset) aiming to achieve. Is there an example for this?
• How would mod:SemanticArtefactCatalogRecord relate to the mod:SemanticArtefact URI? Is this through annotates? Wouldn't the domain/range need to be reversed if so? (read a record annotates an artifact, right?). Can a semantic artefact annotate itself? (e.g., if the metadata is self-contained?)
• Are there any authoritative metadata properties? For example, I do not find dc:creator, but I see there is a http://w3id.org/codo#contractedVirusFrom. I also see there are 2 properties used for versioning. Which ones are the ones needed?

[jonquet]

Dear Daniel, sorry for late reply. Please remember that MOD is work in progress, here are a few elements of responses:

I understand what a mod:SemanticArtefact (SA) is, but not that well what is the "mod:annotates/describes" property (linking it to a Dataset) aiming to achieve. Is there an example for this?

A mod:SemanticArtefact as you see is a subclass of a dcat:Dataset but also a dcat:Resource => The idea here is to consider a semantic artefacts as a "dataset of concepts/classes/properties/individuals" to benefit from DCAT modelisation of the notion of dataset.
But we do need another property to capture also that a specific SA can be used to describe or annotates a dataset. This is the major use of SA with data. For instance:
a mod:SemanticArtefact
a dcat:Dataset
mod:describes/annotates

Can a semantic artefact annotate itself? (e.g., if the metadata is self-contained?)
Why not... not impossible but this is not the reason of existence of the mod:describes/annotates property. I don't really see an example for this.

How would mod:SemanticArtefactCatalogRecord relate to the mod:SemanticArtefact URI? Is this through annotates? Wouldn't the domain/range need to be reversed if so? (read a record annotates an artifact, right?).

A mod:SemanticArtefactCatalogRecord is the object corresponding to the SA but stored/hosted in a specific mod:SemanticArtefactCatalog. It will relate to the main SA with the foaf:primaryTopic property suggested by DCAT.

Are there any authoritative metadata properties? For example, I do not find dc:creator, but I see there is a http://w3id.org/codo#contractedVirusFrom.
Not sure what you mean here. dc:creator is one of the recommended property in MOD2:
https://github.com/sifrproject/MOD-Ontology/blob/master/mod-v2.0_profile.ttl#L1390
Where do you see codo:contractedVirusFrom exactly? This si another ontology developped by @biswanathdutta so I am assuming this will be an error if you would see this in MOD.

I also see there are 2 properties used for versioning. Which ones are the ones needed?
In fact MOD as identified several metadata properties to represent the same things (set of crosswalks). They are all defined in the OWL file generated by @biswanathdutta .. but what you want to look at is the profile here :
https://github.com/sifrproject/MOD-Ontology/blob/master/mod-v2.0_profile.ttl#L433
=> here we have one "recommended" by MOD property (owl:versionInfo) and 6 crosswalks.
Also, in case :

• https://github.com/sifrproject/MOD-Ontology/blob/master/mod-v2.0_profile.ttl#L1607
  owl:PriorVersion is another different property to store the relation between an SA and the previous version of this SA
• https://github.com/sifrproject/MOD-Ontology/blob/master/mod-v2.0_profile.ttl#L1710
  dcterms:hasVersion is a different property used to store the relation between an SA and any other SA before

Hope it helps

[jonquet]

@dgarijo Do you have your responses ? can we close this?

[dgarijo]

Hi @jonquet,
I would keep this issue open until:

• There is an example with annotates (keeping the KGs/datasets a vocabulary annotates is too much IMO, but if you want to have that property I won't fight it).
• There is a way in the documentation to express that one of the properties is the authoritative one. https://github.com/sifrproject/MOD-Ontology/blob/master/mod-v2.0_profile.ttl#L433 is a 404. Otherwise people won't know which one to use. Maybe a better way of doing this is leave in the documentation the main authoritative one and then point to a mapping with all the other ones?
• The typos from external ontos are gone (done)