The second workshop in the series dealt with the criteria to be applied when deciding if a data product must be in the NII and what conditions should adhere to that data once there.
The most common view was that the criteria need to be simple yet reliable and should include a sense of the benefits of a data product being on the NII. From the outset, there was particularly high support to base the selection on two criteria - firstly, the role of a data product as core reference and secondly, whether the data product defines one of the classifications in the NII structure.
This sparked a lively discussion on core reference. The general view was that there is a layer of the NII that should be pure “core reference data” i.e. data that is referenced by other data - essential for data connectivity (location products were requently cited as examples). While there was a strong feeling that this core reference data should be open, many took the pragmatic view that data not being open should not prevent a reference product being in the NII. As such, any closed or unpublished data product appearing on the NII should also provide a clear explanation of how this being closed or unpublished is likely to become an obstacle for using the reference layer. It was further noted that it would be helpful to provide a data owners’ open data plan for those products.
On conditions, it was felt that although having a defined set of quality conditions for NII data was important, the idea of publish first, fix later should still apply to the NII. The group felt that while a prescribed level of quality will have to be achieved by all data ( ensuring accountability to the public) s, not meeting that quality at the start shouldn’t be an obstacle to publication or inclusion.
Meanwhile, timeliness was highlighted as one of the most critical conditions of any NII data If something is said to be published monthly, then it should appear monthly or business will not be inclined to use it.
It was argued that inclusion on the NII list should not be a binary ‘in/out’ and needs to be more flexible. It was agreed, however, that this idea needs more exploring, as a non-binary selection process may make the NII too fluid to be reliable.
A final powerful point made was that effective governance would be essential. The implementation of the NII must be managed by a body or individual with sufficient levels of authority to ensure the integrity of the process.