Create new format-id for science-on-schema.org Dataset in JSON-LD

[datadavev]

Format Metadata

• formatId: science-on-schema.org/Dataset/1.2;ld+json
• formatName: JSON-LD metadata
• formatType: METADATA
• mediaType: application/ld+json
• extension: jsonld

Format description

Dataset resources may be described by schema.org/Dataset markup serialized as JSON-LD in dataset landing pages. In order for DataONE to collate and index such metadata it is necessary to provide a suitable formatId.

Specification / Namespace documentation

• https://science-on-schema.org/
• https://schema.org/Dataset

Checklist

• [x] The format is not a duplicate of another in the list under a different name or identifier
• [x] Format Identifier is unique
• [ ] Format identifier is the commonly-used identifier for the namespace, or the best URI for the namespace, or the best Media type
• [ ] Format identifier is same as the MIME media type if the mime type is specific to only this format (e.g., image/png is specific to one format, whereas text/xml is not specific to one format)
• [ ] Format Name is recognizable and sensible
• [ ] Format includes version info where applicable in formatName and formatId
• [x] formatType is the correct type from the values: DATA, METADATA, or RESOURCE
• [x] MediaType is the most specific MIME media type that applies to the format

Considerations

application/ld+json is likely the most common serialization format, though technically any of the RDF formats may be used.

The formal definition of Dataset is at https://schema.org/Dataset though science-on-schema.org provides numerous recommendations that are necessary for effective use on science data.

Versioning is necessary as recommendations change over time.

[mbjones]

This looks good to me @datadavev . Are you happy with the formatId as proposed? Does anyone else want to weigh in on that before we create a PR?

[mbjones]

My main comment is the formatId is complicated and would likely invite typos.

[csjx]

@mbjones @datadavev - While I think it's a fairly complex formatId string, we have others that are not terribly memorable and prone to typos as well. That said, it's important that we capture the type and version in the formatId and the formatName, and I think this formatId does this pretty concisely. Perhaps the formatName value could be Science on Schema.org Dataset JSON-LD Metadata Version 1.2 or something a bit more verbose. Those are the user-facing descriptions, so I think need to convey the details.

[amoeba]

I'm not sure I see why the ;ld+json is in the formatID. Can we just have it as science-on-schema.org/Dataset/1.2? Otherwise I'm fine with this.

[datadavev]

Because the content can be serialized in multiple formats. This formatId is specific to the JSON-LD serialization.

[amoeba]

Ah, right. Thanks @datadavev.

[taojing2002]

I am using the format id science-on-schema.org/Dataset/1.2;ld+json in the metacat implemenation. If we will change it, it is better to do it right now to avoid more effort to modify code.

[taojing2002]

Removed the version from the format id: science-on-schema.org/Dataset;ld+json