I’m just asking a “daft laddie” question ( http://goo.gl/NDbYsz ) about implementing DCAT catalogues am struggling with practical ways to implement.   The question is about numbers of catalogues and nesting of catalogues, and also about datasets and nesting of datasets.

So, in many organisations, there are many types of information that are published, and some clustering and theming takes place in the publication pathway.  Equally, some datasets are accreting datasets – each year a new dataset is added to an existing dataset.  This is sometimes more than just adding another row of data to a table, it’s more like adding another table to a collection of tables.

Is it satisfactory to have just a single catalogue for the whole organisation, or are (as I think) multiple catalogues easier to work with?  If the latter, how should these catalogues be related to each other (other than indicating that they are published by the same organisation)?  Should there be a "part of" relationship between dcat:Catalog instances?

With respect to datasets a similar scenario exists in accreted datasets.  In some respects each year’s entry should be recorded and be discoverable as an individual dataset. But there should also be a way to ‘nest’ datasets using a "part of" relation.  I think that  this is a gap in the current specification which is something like in SKOS where there is a gap for hierarchical general-specific relations that XKOS etc are now trying to fill.

Component

Category