Consider an article with a DOI. For the CrossRef metadata we could use the DOI as the subject id. If that article is also cited by another article, we could use the DOI as the object id. This means that any attributes of that citation record (e.g., the article title) will be associated with that DOI. Hence we may have two different titles for the same article, one from the source article, one from the citation. This complicates constructing entities from the triple store. We could handle multiple strings, or we could do something like treat titles from the primary source as schema:name and titles from secondary sources as schema:alternateName. This would still leave us with how to handle other properties, such as schema:datePublished.
Consider an article with a DOI. For the CrossRef metadata we could use the DOI as the subject id. If that article is also cited by another article, we could use the DOI as the object id. This means that any attributes of that citation record (e.g., the article title) will be associated with that DOI. Hence we may have two different titles for the same article, one from the source article, one from the citation. This complicates constructing entities from the triple store. We could handle multiple strings, or we could do something like treat titles from the primary source as
schema:nameand titles from secondary sources asschema:alternateName. This would still leave us with how to handle other properties, such asschema:datePublished.