Linked Data mapping

[tkuhn]

Provide all public data of the journal as Linked Data, based on the SPAR ontologies and the Semantic Web journal's Linked Data representations (http://semantic-web-journal.com/sejp/page/semanticWebJournal).

[LEHunter]

The link is bad. The journal itself has a ".net" suffice (http://semantic-web-journal.net ) but I can't find the page being referred to at that site.

[tkuhn]

It seems they are using both domains, but the ".com" one is currently down. This is a snapshot from the Internet Archive: http://web.archive.org/web/20160425100957/http://semantic-web-journal.com/sejp/page/semanticWebJournal

[LEHunter]

OK. Unfortunately, the IA didn't capture the links, so it's a little hard to interpret the SWJ terms. Some of them seem worrisome at first glance, e.g. swj:ucomp-protege-plugin-crowdsourcing-enabled-ontology-engineering

Also, I see that the SPAR ontologies are not aligned with the OBO Foundry, e.g. the Information Artifact Ontology represents many of the same entities as the Document Components Ontology, but I could not find a mapping between them. Further, it seems to me that we should think hard about compatibility with the NIH Data Science Commons ideas. The NIH standards do NOT specify a particular ontology to use, but define "digital object compliance" which is related.

Was there a discussion of the rationale for adopting this approach, and is there an archive of it I could take a look at? Seems to me important to consider this carefully.

[tkuhn]

Some of them seem worrisome at first glance, e.g. swj:ucomp-protege-plugin-crowdsourcing-enabled-ontology-engineering

I think these are identifiers for the papers, with the worrisome looking part just being the words from the paper's title.

Also, I see that the SPAR ontologies are not aligned with the OBO Foundry, e.g. the Information Artifact Ontology represents many of the same entities as the Document Components Ontology, but I could not find a mapping between them.

I was not aware of the Information Artifact Ontology. @essepuntato, what's the relation between SPAR and the ontologies from the OBO Foundry?

Further, it seems to me that we should think hard about compatibility with the NIH Data Science Commons ideas. The NIH standards do NOT specify a particular ontology to use, but define "digital object compliance" which is related.

Yes, good point. It seems that "digital object compliance" and FAIR have the same core.

Was there a discussion of the rationale for adopting this approach, and is there an archive of it I could take a look at? Seems to me important to consider this carefully.

No, it was more a suggestion and not a decision. I wasn't aware of these alternatives to SPAR, so I didn't phrase it as a question. We can have the discussion here.

I thought it's easier to work on this once we have actually some data to model, i.e. once we have some first papers, reviews and decisions.

[essepuntato]

Hi @tkuhn and @LEHunter,

I haven't heard about the IAO before, I apologise. As far as I can see, IAO has a broader scope than DoCO, since it describes other (even foundational, if I can) entities, and some of them are contained in other SPAR Ontologies, such as FaBiO (e.g. when referring to the subclasses of "document"). Thus, there are things that are not specified in DoCO (and FaBiO, and other SPAR Ontologies) that IAO has, and (vice versa) there are other things that IAO does not have that are included in DoCO (and FaBiO, and other SPAR Ontologies).

I think there is margin for a two-way alignment (OBO to SPAR, and SPAR to OBO), since there are several points in common, and it would be good to try to think about it for future releases of both the ontology, I believe. However, I think this should be discussed carefully before starting to implement it, due to the actual foundational models that are used respectively - in SPAR we mainly rely on FRBR for describing documents, while I haven't see such approach used in OBO, at a first (quick) analysis.

In the past, we have actually run an harmonisation among two different ontology project, i.e. SPAR and SWAN, so as to use directly SPAR within the SWAN ontology ecosystem. I don't know if we can used the same approach in OBO (it would be good indeed). Something to discuss indeed.

Maybe, at a first stage, the most secure and quick way for thinking about an alignment is to consider SKOS properties (skos:closeMatch, etc.) for providing such links, without affecting the ontological consistency of both the models.

What do you think?

[micheldumontier]

The other aspect that we need to think about is what does adoption look like for these ontologies for this task. Can we gather some numbers?

[essepuntato]

Hi @micheldumontier,

Difficult to say. In the SPAR website we have some initial insight about the uses of SPAR in different contexts. However, that page should be updated, since there are more project to add - something I would do the next month, hopefully.