search

inab / benchmarking-data-model

OpenEBench Benchmarking Data Model repository
Creative Commons Attribution Share Alike 4.0 International
2 stars 7 forks source link

Should _`email`_ be a required key in `Contacts`? #130

Closed jmfernandez closed 1 year ago

jmfernandez commented 3 years ago

Contact entries are supposed to be there for attribution purposes. So, there could be some cases where who has prepared the entry does not know a valid e-mail address of the Contact. Even worse, there could be some corner cases, like people leaving science or deceased researchers where the e-mail is either unknown or not meaningful.

jlgelpi commented 3 years ago

The concept of "contact" implies that shoud be a way to actually contact him/her. Retired/deceased people cannot "contacts" for communities or tools. They can be "authors" or contributors but never contacts

jmfernandez commented 2 years ago

Comment https://github.com/inab/benchmarking-data-model/issues/129#issuecomment-1178871829 is related to this issue. The whole benchmarking data model should be analyzed, so all the relations to Contact are truly to contacts or they should be pointing to a new Contributor concept.

jmfernandez commented 1 year ago

Conclusion: Contact entries always have an ORCID, either real (for real people) or a fake ORCID (computed from e-mail address using sha256-48, integrating it in an ISNI starting with 9). So, a new restriction is that Contact has a single e-mail address instead of a list of them.

jmfernandez commented 1 year ago

Comment #129 (comment) is related to this issue. The whole benchmarking data model should be analyzed, so all the relations to Contact are truly to contacts or they should be pointing to a new Contributor concept.

So, if we want to restrict the relationships to real people (aka Contributor), all the foreign keys should also include a restriction about the ORCID, like "pattern": "^orcid:0"