2024-11/12 GreenData4All Inception Report - Questions to INSPIRE MIG

• Does this question refer only to legal provisions (in the IRs) or also the technical provisions (in the TGs and Good Practice documents)?
• Does this question refer only to the provisions on data interoperability or also on the technical provisions on network services and metadata?
• For data interoperability, there will always be the need for some processing (ETL) steps before being able to use data from different sources in an application or integrate it in a (pan-European/cross-border) data set. The main question here is how this necessary transformation is performed – by requiring all providers to transform their data to a common (exchange) model (the current obligation) upfront, from which a user/application developer can then transform the data to their application needs, or whether to put more burden for the transformation from the source format (data “as-is”) on the user/application developer.
• It should be considered that even with the current approach, harmonised INSPIRE data can usually not be used in an application/integrated in a dataset “off-the-shelf” (i.e. with minimum extra effort), but that there is still considerable effort involved at the application developer side, as illustrated e.g. by the work of ESTAT on building pan-EU data sets.
• For any transformation to a target application/data set, the user/application developer will need as a minimum a clear documentation of the structure and semantics of the data that is provided (ideally also in English and/or some machine-readable format).
• Also, the use of well-established vocabularies for category-based attributes and common approaches for the encoding of identifiers and coordinate information is crucial.
• BGR: The current GML format poses challenges in processing  therefore, alternative formats like Geopackage, which are more common and interoperable, could be more suitable for data exchange.

Responses (Brandenburg):

• Simplification of INSPIRE data structures: Creation and central provision of mandatory core data models in simplified data structures to ensure user-friendliness and applicability. The core data models should based on the existing complex INSPIRE data models. The complex data models remain in place.
• Scalability: Differentiation between local, regional, national, and international needs in the creation of core data models. It has to be ensured that the data structures are uniformly designed across the different (administrative) levels to enable data aggregation across levels.
• Specification: A clear position from the EU on whether national datasets are required for a given data theme.
• Mandatory implementation: All responsible data-holding entities are obliged to provide INSPIRE geodata in a uniform data structure and, if necessary, retroactively align existing data with the standardized structures to ensure comprehensive data coverage for each country.
• Multilingualism: Mandatory implementation of multilingualism in the creation of data schemas and data content (e.g. code lists) (for example, English is mandatory, while the national language is non-obligatory).
• Validation: Provision of test classes and validators to verify the accuracy of data schemas and data content (e.g., code lists).
• International standards: Binding requirements and mandatory implementation of international standards (e.g., OGC, ISO, W3C). The following requirements have to be considered:
• Provision of metadata (ISO, DCAT-AP, etc.)
• Machine-readable formats (GML, GeoJSON)
• Open licenses (definitive list; CC-0, CC-BY, Data License Germany, etc.)
• Provision through standardized interfaces (OGC API, WFS, WCS, Sensor Things API, etc.)
• Bulk download
• Resources and access: Provision of the necessary resources to the national point of contact to collect and centrally publish uniform geodata. Additionally, the EU should formulate requirements for the national points of contact to determine the appropriate resource needs.
• Sanctions: Sanctioning in case of non-compliance with the binding requirements.

 Optional:

• Best practice implementations / Example implementations before publication / resolution, as a guide for the affected entities.

• What exactly is meant by “abandoning the one-size-fits-all approach to harmonization of data”? Do you mean the obligation to transform all data sets under the scope of INSPIRE to the defined common data models upfront? What would this obligation be replaced with? For example, simplification of obligations or a reduction of the burden on data providers could also mean a common approach/common requirements for all data sets, but with fewer ex-ante obligations or focusing obligations on documentation of “as-is” data rather than transformation.
• Removing the obligation to harmonise all data upfront could potentially also remove the duplicate implementations in some (most?) MS, where “as is” data are used in national applications and use cases,  that is done only to meet the INSPIRE requirements, i.e. the same “as-is” data that is available in the national SDI would also be made available under INSPIRE.
• BGR: Abandoning the principle of one size fit all would contradict the principle of interoperability and harmonisation of spatial data across political boundaries.

Responses (Brandenburg):
If this approach were abandoned, it would

• be impossible to achieve interoperability of data content and integrate data from various sources.
• make a cross-border/cross-jurisdictional evaluation and analysis of geodata impossible, preventing answers to international questions.
• fail to fulfill the objectives of INSPIRE.

 Notes/Follow-up questions:

• It would be helpful to know which questions arise from the EU that have to be answered by using the distributed data.
• If the goal is to unlock economic potential, it would also be interesting to know the specific requirements for data content.
• It should be explored which other approaches to data harmonization are necessary and what is needed therefore. The use of IoT, ontologies in combination with ML, LLM, and AI could potentially be a target.

• The fewer technical provisions for upfront harmonisation exist, the more labour-intensive will be the transformation to a target application/data set. Therefore, if requirements for upfront harmonisation were to be reduced/abolished, sufficient documentation of data structures and semantics (e.g. including a mapping to well-established vocabularies) would be crucial.
• BGR: It may result in interoperability challenges if INSPIRE does not mandate mapping these vocabularies to accepted standards as part of the data-sharing process

Responses (Brandenburg):

• Significant regional/national differences in the required data content could be expected.
• In the process of standardization, relevant data content might get lost. Therefore, local, regional, national, and international differentiation have to be considered when creating core data models (see above).
• Monetary and personnel efforts might complicate the implementation of simplified INSPIRE data structures in member states.

• Dropping the obligation to harmonise all data sets within the scope of the INSPIRE Directive upfront would mean considerable cost reductions for data providers, especially if duplicate implementations could be avoided through this.
• Of course, additional costs for implementing alternative measures (e.g. an obligation for thorough documentation) would need to be considered when calculating cost reductions. 

Responses (Brandenburg):

• Implementing simplified INSPIRE data structures and mandatory requirements could initially incur costs.
• However, cost reductions could be expected after implementation, as the aggregation of cross-border, simplified, and harmonized INSPIRE data will require less time. Time-intensive preparation of INSPIRE data could be eliminated.

• Some OASC MIMs (e.g. MIM-2, MIM-7) could be good sources for inspiration for alternative obligations, but they seem to still be at an early stage in their development, so should not be taken up 1:1.

Responses (Brandenburg):

• The INSPIRE Directive should adopt the approach of MIMs (Minimal Interoperability Mechanisms) and continue this by using international standards. Simplifying and converting the previously complex INSPIRE data structures into flat data structures is recommended in this context.
• We support the introduction of MIMs because they focus on data relevant and specific subject-oriented questions.
• It is questionable whether all institutions responsible for this data are involved in the process and whether they are aware of the specific MIMs.

Notes/Follow-up questions:

• Will the MIMs be adopted as mandatory?
• MIMs have to be designed by considering future developments and requirements.

• The question is not really clear. The impact of what precisely?
• Generally, the overall future vision for the provision, sharing and use of geospatial and environmental data should be defined first. Then it should be discussed how this vision can be achieved and which (parts of which) existing legislation should be changed how and which aspects can be covered through soft measures (e.g. the Green Deal dataspace). 
• GBR: Harmonizing the requirements of HDV, OPD, and Green Data space would be logical, as differing standards could lead to increased effort and complexity

Responses (Brandenburg):

• The initiative of the HVD Directive is supported as it directly affects all entities holding (geo)data and requests for the technical provision of relevant data via standardized web services.
• However, there are concerns that the technical implementation for the labeling of HVD metadata will be realized differently within the EU member states. In addition, differences in license information are expected. The provision via difficult to handle standards (e.g. Atom) is also viewed critically in this context.
• Therefore, specific technical guidance and requirements for the interoperable implementation of these initiatives within the member states of the EU would be recommended.
• Furthermore, a clear regulation of responsibilities would be welcomed stating that the national contact points for the INSPIRE Directive are also responsible for coordinating HVD initiatives concerning geospatial data, as there are significant overlaps between the two legislative initiatives.

Who should be responsible for the governance of

• collecting/assessing the requirements,
• identification of core datasets,
• endorsment of technical solutions.

• This strongly depends on how many details are regulated in the Directive/Implementing Acts and how much is left to guidance or standardisation. However, it could be useful to set up a process for collecting/assessing user requirements with a central point of contact at the EC that receives requirements from the community. We have already established such a process in Germany (https://bedarf.gdi-de.org).  
• What is meant here by “identification core data sets”? The definition of what data sets shall be in scope of the revised Directive? Or who in each MS identifies which of the data sets comply with that definition?
• The development of technical solution could be left to technical communities (e.g. in the GDDS), projects or standardisation bodies. There should still be a step of formal endorsement of recommended solutions (similar to the current Good Practice Process).

Responses (Brandenburg):

• Identification and collection of datasets:
  • The national contact point (in Germany, LG GDI-DE with Kst. GDI-DE) assumes steering and coordination.
  • Central entities at the national level (e.g., federal authorities) should uniformly identify relevant datasets within the nations of the EU and centrally collect them across federal structures.
  • These central entities (in Germany, federal authorities) should actively approach data-holding organizations and verify that all required data has been delivered to achieve spatial data coverage. Additionally, technical assistance should be provided by these entities.
  • Furthermore, a national institution should compile all identified and collected datasets into a web application and (in parallel) provide them to enable centralized access to all INSPIRE content.
  • Moreover, all INSPIRE data will continue to be searchable and available in the EU geoportal.
• Technical solutions:
  • Numerous technical solutions already exist for the implementation of INSPIRE data, INSPIRE services, and the collection of INSPIRE metadata. Many technologies (web applications) are also available for collecting and centrally providing INSPIRE content. Central institutions (e.g., federal authorities) should provide the national and standardized datasets in these web applications and take responsibility for their maintenance.

• The current process for the management of INSPIRE Good Practices and Artefacts (Technical Guidance Documents, XML schema, UML-models, INSPIRE Registry content) seems to work well.

Responses (Brandenburg):

• The organization depends on the various governmental structures and hierarchical levels within the member states. In any case, a national contact point has to be designated to centrally publish all INSPIRE data.
• Simply harvesting metadata should be reconsidered, as this has so far produced insufficient results. A central collection and provision of national INSPIRE datasets, along with a coordinated update cycle, should be considered in this context.
• The possibilities for feedback at the EU level are currently restrictive and limited. Particularly, feedback on technical questions often occurs with significant delays, leading to individual technical solutions, which handicap the expansion of interoperability.

• The development should be community-driven (and –paid), the endorsement and management process should be covered by the EC.

Responses (Brandenburg):

• It is a cross-border interoperability project of the EU. Therefore, the EU should provide funding. Consequently, member states would contribute to the costs based on an appropriate distribution key.
• Funding via EU funding programs or EU budget resources (e.g., infrastructure programs) would be conceivable and should include the support for mandatory tasks. Additionally, funding should include personnel and technical resources as well as training. Furthermore, the collection of datasets that are not yet available but required under INSPIRE should be financed to achieve full data coverage.