In the German Data Portal http://govdata.de we rely on the fact that there is exactly ONE license per dataset and that this license covers all the resources or distributions in the dataset. We would run into trouble we were to aggregate portals that have multiple licenses in a dasaset. If different licenses apply, we suggest splitting into multiple datasets.
On the political axis I fear that this can lead to patterns like "we license the jpeg openly, but the vector verision commercially."
Does any one face a similar problem?
Can we at least encourage resctriction to one licenses per dataset.
Component
DocumentationCategory
improvement
Login or
create an account to comment.
Comments
In fact current draft (v0.03) already restricts license with max. cardinality = 1
This issue raised here concerns the possible situation that there is more than one Distribution for the Dataset, and that these Distributions have different licences.
In the DCAT model, the property dct:license is associated with a Distribution. Govdata.de works with a different model where the licence is associated with the Dataset.
This incompatibility fo modelling is difficult to solve.
On a practical level, there is no problem if there is only one distribution or if all distributions of a dataset are available under the same licence (which may account for the majority of cases); if there are multiple distributions with different licences, an aggregator may need to select one distribution, e.g. the one with the "most open" licence, and forget about the rest.
That is a limitation of govdata.de. There are two ways to solve this for govdata.de: