curious if anyone has some thoughts on the following: Traditionally (especially in the geospatial community) we are accustomed to manually create metadata e.g. according to iso profile 19115 to comply with inspire data publishing guidelines. This metadata is than published via OGC CSW (Catalog Service for the Web) services, pointing to the respective data services. In modernizing our whole data ecosystem I want to introduce a product like datahub to automate data discovery, facilitate insight in lineage etc. But we now have this big 'cultureclash' between the 'manually inclined' colleagues vs the 'automation inclined' people. And apart from culture, there's still lots of value in the manual curated documentation that we want to augment the datacatalog with. So with the hackathon I demonstrated one way of bridging the RDF/Shacl world with datahub, and that might certainly be a way to bridge those geospatial metadata challenges as well. But I'm also interested if anyone has some other insights to share on this. e.g. templating the documentation tab, so it shows up the iso 19115 fields or whatever. Thoughts anyone?

Would this metadata typically apply at the dataset level?

So, technically I think it can be on both a table and a dataset level. The most 'visible' manifestation is usually on a dataset level indeed I think. Which then typically ends up being harvested from CKAN or the likes.