As open data user I need NISV to provide proper linked open data, so I can refer to media objects and terms.

[wmelder]

Task: Define steps to take with the insights coming from Wikimedia scan.

(Maybe create a new EPIC for the open data lab that describes how we can improve current workflows using structured data for commons (with Wikidata).)

Currently, we have two uploads of our public domain stuff. One to Open Beelden, one to Wikimedia Commons. The publication metdata of the Open Beelden platform is not really LOD. It is DCMI plus ccRel, but not in RDF format. This could be upgraded to at least these standards expressed as RDF. But, more important shared terms are now (literally) copies of labels from the source system (DAAN) to Open Beelden. This process should be revised. A transformation of the source metadata (preferably to a open standard like DCMI or schema.org and ccRel) would be the goal. References to stable (cool) identifiers is very convenient, but the use of the NISV persistent identifiers might well be good enough already. The current software for the OpenBeelden platform is not likely migrating to a linked data platform. So, alternatively, we might consider to revoke the open beelden publications (or discourage the use of them) and publish the Public Domain really as Linked Open Data (maybe using an RDF store with SPARQL endpoint and taking schema.org as standard for findability into consideration). From there, uploads to other platforms is more structured (Commons and Wikidata), refererence back to the item sources is more consistent (the Open Beelden item IS the same item as the DAAN item), and linking and enriching the metadata for these resources using other platforms is more opportune (enrichments coming form Wikimedia tools, after inspection and approval, can be (automatically) ingested by DAAN).

Ook IIIF hierin meenemen... Met name de IIIF AV working group in de gaten houden..

NOTE: IIIF lijkt JSON-LD als default serialisatie te ondersteunen. Da's mooi, want da's RDF, dus eerste stap naar meer interoperabiliteit!

'Shared Canvases', een data model waar ik nog nooit van gehoord heb: https://iiif.io/model/shared-canvas/1.0/

Zie IIIF en Wikimedia

TODO

• [ ] find out what is currently the practice in publishing open beelden items
• ask ruurd? or is it the same workflow as below?
• [x] find out what is the current practice for upload to Wikimedia Commons
• https://www.beeldengeluid.nl/en/knowledge/blog/first-batch-upload-wikimedia-commons-using-gwtoolset
• https://gitlab.beeldengeluid.nl/o-team/openbeelden-uploader
• [ ] compare current publication format (OI) with Wikimedia Commons metadata
• [ ] describe the current persistent identifiers, resolver and landing page
• [ ] show the use of shared terms and the points for improvements
• [ ] describe a future software architecture model serving the information needs
• [ ] a decision needs to be taken how to proceed (for example whether we choose a temporary solution to work with)

Acceptance

• [ ] the target group owner needs to be consulted
• [ ] user groups need to be consulted

[wmelder]

add to next sprint

[wmelder]

• [ ] Compile a document (maybe publish on the github) that describes the current solution.

[wmelder]

eigenlijk te groot voor issue... is meer epic