GML issues

[anitagraser]

GML is currently recommended for vector data, see:

• Guideline_CITS_Geospatial_v_1_0.pdf
• Guideline_CITS_Geospatial_Appendix_1_v1.pdf

However, GML (and also GeoJSON) files tend to get very large for larger datasets (such as, for example, address data) which makes it very difficult to deal with (read/write) the files.

Data producers therefore lean towards using the GeoPackage or even just CSV with WKT.

[anitagraser]

GeoPackage is, for example, in use at the German GDI-DE https://wiki.gdi-de.org/display/Arch/Verzeichnis+GDI-DE+Standards

[GregorZavrsnik]

Thank you for this proposal.

We welcome all suggestions for new long-term preservation formats.

GML has its benefits and its weaknesses, however, it suits all the criteria for the long-term preservation of the format, although it cannot be quickly used in some GIS applications. But that is not the main criterion for a long-term preservation format.
The main criteria are Openness (GML is an ISO Standard), Adoption (It is adopted OGC Standard), Self-documentation, Robustness, and Dependencies. GML is the long-term format of choice for some archives because it suits most of these requirements. That is why it was added as an example in the Appendix as an example of a long-term preservation format profile.

But I agree, that the CITS Geospatial is also intended for organizations, that are not official archives and should therefore also support formats, that are better suited for their purposes.

So we depend on the community to contribute more Long-term preservation format profiles for their formats of choice, based on the provided examples.

[GregorZavrsnik]

We aim to deliver more examples of long term preservation format profiles for other formats in the future releases