Distinguish MeataDataEntries semantically

[baffelli]

Description:

Currently, metadata entries in the RO-Crate Interoperability Profile are only linked to their type via rdf:type that SHOULD reference a rdfs:Class, making it difficult to distinguish between schema/type definitions and actual metadata entries during parsing and reconstruction.

Proposal:

Introduce a dedicated predicate (e.g., profile:metadataEntryOfType) to explicitly link each metadata entry to its schema type.
Emit metadata entries with a special rdf:type type (e.g., profile:MetadataEntry) to clearly identify them as metadata, not schema/type definitions.

Benefits:

• Unambiguous identification of metadata entries.
• Cleaner, more robust parsing and reconstruction.
• Aligns with best practices for schema/instance separation.

[AndreasMeier12]

Okay, so to make sure I understand this correctly, a metadata entry of a Publication would look like this.

{
"@id": "1234",
"@type": "profile:MetadataEntry",
"profile:metadataEntryOfType": "Publication"
}

I omitted all properties and schema definition for this example.
Is this understanding correct?

I see the benefits, this is definitely worth a closer look.

[baffelli]

Hello Andreas, That's exactly how I imagined it to be. I believe it would reduce ambiguity and make ser/de easier to implement

…

On Mon, Aug 25, 2025, 12:47 AndreasMeier12 ***@***.***> wrote: *AndreasMeier12* left a comment (researchobjectschema/ro-crate-interoperability-profile#1) <#1 (comment)> Okay, so to make sure I understand this correctly, a metadata entry of a Publication would look like this. { ***@***.***": "1234", ***@***.***": "profile:MetadataEntry", "profile:metadataEntryOfType": "Publication" } I omitted all properties and schema definition for this example. Is this understanding correct? I see the benefits, this is definitely worth a closer look. — Reply to this email directly, view it on GitHub <#1 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AF7KCDIWY3MDEVX4DO2MVKT3PLSVRAVCNFSM6AAAAACEHNMJJWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTEMJZG42TIMZZHE> . You are receiving this because you were assigned.Message ID: <researchobjectschema/ro-crate-interoperability-profile/issues/1/3219754399 @github.com>

[AndreasMeier12]

From my side, this can go into the next version of the specification.

[AndreasMeier12]

After discussing this in today's meeting, we found a simpler way of including this information.

Current

{
"@id": "1234",
"@type": "Publication"
}

Simone Proposal

{
"@id": "1234",
"@type": "profile:MetadataEntry",
"profile:metadataEntryOfType": "Publication"
}

Juan - Andreas Proposal

{
"@id": "1234",
"@type": "Publication",
"schemapp:kind" : "meta"
}

{
"@id": "Publication",
"@type": "rdfs:Class",
"schemapp:kind" : "class"
}

[minottic]

after the meeting last week (sorry!), I think this is fine, as long as it does not have consequences on the parsing of crates that do not use custom rdf classes. For example, what happens if I use a native schema.org CreativeWork? Should I just omit the schemapp:kind" : "class" and only add the "schemapp:kind" : "meta" when creating the metadata?

{
"@id": "1234",
"@type": "CreativeWork",
"schemapp:kind" : "meta"
}

[baffelli]

I think so. However, I would prefer something more precise:

• Classes are defined with rdf:type rdfs:Class (optionally rdfs:subClassOf …).

• Metadata entries are not rdf:Class and do not use an existing class as their type. Instead we mark them with a dedicated type (e.g. profile:MetadataEntry) and link them to the schema/type they describe via a dedicated predicate (e.g. profile:entryOfType).

This cleanly separates schema layer from metadata/instance layer.

To do this properly, we should have a small ontology and publish it so that we can include it in our prefix. E.g:

@prefix profile: <https://example.org/schemapp#> .
@prefix rdfs:    <http://www.w3.org/2000/01/rdf-schema#> .

schemapp:MetadataEntry a rdfs:Class .
schemapp:entryOfType   a rdf:Property ;
  rdfs:domain profile:MetadataEntry ;
  rdfs:range  rdfs:Class .