Conclusions and Outlook
First of all, we (the project team) would like to thank the INSPIRE community for the incredible amount of support and input that you provided! You supported us as members of the Expert and Advisory Board, through filling out the survey, or through encouraging us to go forward. Based on this input, we see this work to be highly relevant.
From the survey we quickly learned that there is a real need to define an extension methodology, even among the mostly expert audience that provided responses. On the implementation side, the main challenges lie with system integration and data harmonisation.
The methodology covers a lot of ground with its seven phases, but doesn't go into all the nitty gritty details of conceptual modelling. This is where more generic resources on Object-Oriented Design and UML will be valuable additions. Check out these links for more on these topics:
- Wikipedia: Object-oriented analysis and design
- Wikipedia: Types of Relationships in UML Class Diagrams
- ISO-TC211: Best practices for diagram design
We've collected a lot of valuable insights on the impacts that different conceptual modelling techniques have when you build a dat model that is in compliance with the INSPIRE data specifications, such as ruling out bidirectional associations. Building out each example and iterating over it multiple times has enabled us to find most challenges.
There are still a lot of open questions about additional patterns. We see some emerging that would be highly useful, such as Redefine, but the tooling is not complete enough to recommend their usage. We will continue to develop the tools we used to analyse model extensions, and there are more than 20 more model extensions out there that we haven't yet published.
Here are some of our concrete plans:
- Continue to develop this site into an official registry for extensions
- Add a simple self-registration form for more extensions instead of the survey
- Improve the statistics, model links, and pattern identification using concrete models, by using automated analysis
- Make sure we get access to the actual model for as many extensions as possible
- Include extensions in the interactive data specifications find your scope tool
- Add sandboxes for interactive work with data models to tutorial project
- Add interactive model exploration tools to each model extension entry for which we have a formal model (XSD, UML)
- Add a section on UML modelling requirements and best practices, such as UML model presentation and the integration of all relevant ISO 191xx requirements
We plan to continue this work, so please continue to submit contributions!