ESIPFed / science-on-schema.org

science-on-schema.org - providing guidance for publishing schema.org as JSON-LD for the sciences
Apache License 2.0
109 stars 31 forks source link

SOSO guidance should provide an identifier for metadata that conforms to the spec #169

Open smrgeoinfo opened 3 years ago

smrgeoinfo commented 3 years ago

It would be useful for client applications to know that a given metadata record conforms to the science on schema.org recommendations. Currently 'conformsTo' is a schema.org property under discussion, but dct:conformsTo is already widely used. Usage would be like:

 {
  "@context": {
    "@vocab": "https://schema.org/"
  },
  "@id": "metadata:lod.example-data-repository.org/id/dataset/3300",
  "dcterms:conformsTo":"https://github.com/ESIPFed/science-on-schema.org/1.2.0",
  "identifier":"https://doi.org/10.1234/1234567890",

In the DataSet.md guidance document, we'd need to add some introductory text stating that "To identify a metadata record as conforming to the recommendations in this document, use the URI 'https://github.com/ESIPFed/science-on-schema.org/1.2.0' as in the example above.

charlesvardeman commented 3 years ago

There is also the Profiles Vocabulary from the W3C Dataset Exchange Working Group. The Profiles Vocabulary uses dcterms:conformsTo connecting an rdf:Resource to a dcterms:Standard. Profile is a subclass of dcterms:Standard so that would work for science-on-schema.org except I think it needs to become a URI.

 { 
"@context": {
    "@vocab": "https://schema.org/"
  },
  "@id": "metadata:lod.example-data-repository.org/id/dataset/3300",
  "dcterms:conformsTo":[{"@id":"https://github.com/ESIPFed/science-on-schema.org/1.2.0"}],
  "identifier":"https://doi.org/10.1234/1234567890"
}

That would allow science-on-schema.org the option to create a dereferenceable identifier in the future where

<https://github.com/ESIPFed/science-on-schema.org>  prof:isProfileOf <https://schema.org/> .

I think science-on-schema.org meets the definition of a prof:Profile, I.e., "A specification that constrains, extends, combines, or provides guidance or explanation about the usage of other specifications."

smrgeoinfo commented 3 years ago

@charlesvardeman we're on the same page. What we need is that URI for the SOSO schema.org profile. It needs to be explictly stated in the guidance document. I think 'https://github.com/ESIPFed/science-on-schema.org/1.2.0' would be a bad choice for the profile URI. It should at least be a github tag. See also #170 for question as to whether the value should be a LOD URI, or just a URI string. What are your thoughts about that?

smrgeoinfo commented 2 years ago

I see that DataOne is using something like this to identify metadata conforming to SOSO 'profile': "formatId": "science-on-schema.org/Dataset;ld+json"

mbjones commented 2 years ago

Yes, that comes from our format vocabulary that we use for typing METADATA and DATA format types consistently across the federation:

https://github.com/DataONEorg/object-formats/