Skip to main content

Data Provider side

Draft – under review

Part of the Stack documentation (overview). Not yet endorsed by NDE.

Data management and Publication are run by Data Providers. Data management is DERA’s Beheer (“het in stand houden van erfgoedinformatie en het verrijken daarvan met aanvullende informatie”): source-internal, and out of scope here. Publication is the source-side externalisation that turns internal data into something the network can consume. The function of each layer is DERA’s to define; this chapter references DERA for that and documents the engineering.

Publication Layer​

Data Providers make their data available via standardised channels so Service Platforms, consumers, and network services can discover, fetch, and process it.

This layer is DERA’s Publicatie: “zorgdragen dat erfgoedinformatie toegankelijk wordt gemaakt voor afnemers”. Unlike the Data Layer, it sits inside DERA’s existing scope, so the terms below are shared rather than Stack-specific.

Patterns applied at this layer​

The Publication Layer is the source-side of several cross-cutting patterns defined in the Patterns chapter:

  • SCHEMA-AP-NDE-first: the inter-layer data contract; every Data Provider publishes at least SCHEMA-AP-NDE.
  • Parallel Data Models: publish in SCHEMA-AP-NDE and, where relevant, in one or more domain models (Linked Art, RiC-O, RDA, EDM) alongside.
  • Enrichments-as-Datasets: when a Service Platform publishes its own enrichments back into the Dataset Register, those enrichments enter this layer as new published datasets.

Standards used at this layer​

The Publication Layer is built on standards published outside this Stack documentation. Each links to its canonical source.

  • SCHEMA-AP-NDE: the NDE-Schema.org-toepassingsprofiel. Network minimum for heritage data publication.
  • Linked Open Data: dataset publication as Turtle / N-Triples / JSON-LD dump or SPARQL endpoint. See Publish linked data.
  • LDES: incremental change feed. Currently the most mature mechanism for sub-dataset-granularity change detection.
  • IIIF: Image API + Presentation API for image-based heritage. See Publish IIIF.
  • DCAT-AP 3.0 / Schema.org Dataset – the dataset-description vocabularies the Dataset Register accepts. Data Providers publish descriptions in either; Schema.org Dataset is the more common choice in practice, DCAT-AP is the EU-catalog-interoperability standard.

Network services consumed at this layer​

Software at this layer​

The Stack builds no software of its own here; instead this layer is filled by NDE-compatible products – third-party software that is part of the ecosystem by being NDE-compatible, not by being provided by the Stack. One product typically covers both Data management and Publication.

Omeka-S is one such product. Van data naar dienst describes it as “een open source, semantisch collectiebeheersysteem, waaraan je verschillende modules kunt toevoegen om het systeem NDE-compatibel te maken” – so the standards above are met by adding modules, not by replacing the system. At the Gouda Tijdmachine it “fungeert als het collectiebeheersysteem” and “ook als publicatieomgeving van de collectiedata”, publishing a dump and a SPARQL endpoint. Two modules connect it to the network services above: NdeTermennetwerk for linking descriptions to terms, and LinkedDataSets for registering datasets in the Dataset Register.

What a collection management system must do to be NDE-compatible is specified in Requirements, not here.