The increasing need for geospatial information demands for well-organised management among all levels of society. A Spatial Data Infrastructure (SDI) is a multidisciplinary and dynamic instrument that facilitates access and sharing of geospatial information. The current trend towards open data initiatives is influencing the development of these infrastructures. In order to examine this effect, this article addresses the following question: what is the current state of SDI openness of four best practice open data countries Canada, The Netherlands, Australia and Brazil, and how do they compare? The question is answered through a qualitative literature study and the application of a newly developed Open SDI Assessment Framework to the countries. The Netherlands and Canada show a high performance on all assessment dimensions; data discovery, data access and data properties. Australia and Brazil show a poor open SDI performance, as they could not meet the requirements set for the assessed datasets. General conclusions of the assessment are that data is currently fragmented and scattered among the web in all four countries, which strongly negatively influences the user experience. It is crucial that a strict legal framework is embedded in a country, which ensures that current SDI objectives and propositions regarding an user-centred approach and open data availability are achieved.

In this paper we present an open and flexible approach for the standardisation of 3D geographical data, describing our physical environment in such a way that it can serve different applications. The aim of our approach is to keep the standard as simple as possible so that implementation in different software is straightforward and the reuse of once collected 3D data in different domains is optimally supported. Therefore, we propose to model the semantics of real-world objects independent from their application and we distinguish between the conceptual model and encoding. The result is a 3-layer approach, in which the first layer contains the conceptual model: the object types with their definitions and properties. This layer reuses definitions of various existing models (national and international) as much as possible. The second layer contains the modelling constraints: the set of rules that define how the objects from the conceptual model are represented in 3D as needed for a specific context or application. This second layer contains additional (3D) requirements to standardise the 3D representations of the objects. The third layer contains encoding profiles, thus specifying how different formats can best be encoded; these formats could be JSON or XML/GML.

In this paper we motivate and describe our approach. For a small area we have developed a prototype that implements the 3 different layers. The prototype shows how the approach can be implemented for one specific application and additionally it provides insight into further development.