The first data source type should not be SPARQL endpoint, as this is the least usual use case. This is suitable maybe for us, but a regular user usually has a file, or a file on the web.
Since empty data sources are ignored anyway, I think there could be 3 datasources from the start:
The first data source type should not be SPARQL endpoint, as this is the least usual use case. This is suitable maybe for us, but a regular user usually has a file, or a file on the web.
Since empty data sources are ignored anyway, I think there could be 3 datasources from the start:
and then the "add more" buttons.