Marking a dcat:Dataset as deprecated (Discontinued)

[ChLaaboudi]

Description: DCAT-AP currently lacks a formally defined mechanism to mark a dcat:Dataset as deprecated. While the property adms:status (with lifecycle status values such as "Deprecated") is defined for dcat:Distribution and dcat:CatalogRecord, it is not available for dcat:Dataset itself. This is a known gap in the specification.

Proposed Solution:
To address this, we should define a property at the dcat:Dataset level, with two potential approaches:

1. Extend adms:status to apply to dcat:Dataset. The object is a dataset-status (Codelist)
   Codelist Reference: The EU Vocabularies dataset-status codelist includes a relevant value for deprecated datasets:
   Withdrawn: http://publications.europa.eu/resource/authority/dataset-status/WITHDRAWN-
2. Use owl:deprecated (with a boolean value of TRUE).
   Best pratices for linking deprecated and replacing datasets

Best practice
To maintain data provenance, deprecated datasets should explicitely reference their replacements (and vice-versa) with the properties dct:replaces/dct:isReplacedBy

[bertvannuffelen]

This is (mostly) adressed at the DCAT level. https://w3c.github.io/dxwg/dcat/#life-cycle. So any profile can just apply this approach. It is a variant of the proposal of @ChLaaboudi.

The sole question resides where to make the decision for the used codelist. Is that a DCAT, DCAT-AP or STAT DCAT-AP or left to the users to choose.

[jakubklimek]

I think, given the EU NAL available, it will not be a DCAT decision. It is also not specific to Statistics. To me, it seems like a DCAT-AP decision, coming from the StatDCAT-AP community.

[ChLaaboudi]

Approach

• Update Request addressed from StatDCAT-AP to DCAT-AP

1. Extend adms:status to dcat:Dataset
2. Link with EU Vocabulary dataset-status codelist (label = Withdrawn)
   Usage note: Recommended or Mandatory? (To decide on)

Best practice

Preserve provenance by linking versions with dct:replaces / dct:isReplacedBy

[jakubklimek]

Related GeoDCAT-AP issue supporting the fact that this is more of a DCAT-AP level decision.

[jakubklimek]

I think there is an agreement on the property and codelist addition. The only thing left is whether we introduce it in StatDCAT-AP now as profile-specific, wait for it to be included in DCAT-AP (SEMICeu/DCAT-AP#242) and then link it in a future release, or we wait for it to be included in DCAT-AP first, and introduce it as reused from DCAT-AP later. @bertvannuffelen ?

[jakubklimek]

In the draft:

• Added Dataset.status
• Added corresponding MUST Controlled Vocabulary usage expectation and link