INSPIRE Thematic Clusters

Dear Jordi, and all

I'd like to provide additional information on the CIS 1.1 Coverage model, which allows other encoding than the GML only and also additional Coverage classes as regards the ones in CIS1.0 (= GMLCOV) that is being adopted by ISO TC211 as 19123-2 Coverage Implementation Schema (CIS 1.0, based on existing ISO 19123).

This new CIS 1.1 Coverage Implementation schema has been standardized by the OGC on basis of 3 encodings in GML, JSON and RDF, and are based on a new Coverage model (whose UML diagrams are provided in html, or other document formats such as pdf or MS word), which is different from the existing OGC AS topic 6 / ISO 19123:2005 (for which a new project has been started). Figure 16 provides a view of the integration of the previous GridCoverage, RectifiedGridCoverage and ReferenceableGridCoverage (coming from ISO 19123 and GMLCOV, and widely used in INSPIRE data specifications for Coverage-based data), and the parallel emergent (as of CIS 1.1) GeneralGridCoverage and GeneralGrid classes. This new UML Coverage model has not been provided to the OGC nor TC211 in order to secure its consistency (both internal and external with other relevant TC211 standards - such as GFM, CRS, ...). This effort has not been done (up to now) because OGC worked on a implementation standard (CIS 1.1) and UML is only informative. The validation of the new UML model has not been done by OGC (under some AS Topic 6), and not by TC211, as of the revision of 19123-1 project, nor integration in the TC211 HMMG model. Consequently, this CIS 1.1 UML model can't presently be considered as standardized, or the people doing so would do it under their own responsibility.

As (for the GML encoding) the CIS1.1 schema (based on this new Coverage UML model) as updated/changed some Coverage concepts, the name of classes / elements and even the encoding of the coordinates, the "backward compatibility" has been achieved (in order to solve comments sent during the balloting period by IGN, DGIWG, as well as other stakeholders) by inclusion of CIS1.0 into CIS1.1 (see requirement #1 and conformance clause. In fact, the new CIS1.1 schema is not - by itself - backward compatible with CIS 1.0 schema, and the specifications provides no mapping of CIS1.0 elements to CIS1.1.

Such a mechanism for backward is acceptable in the XML domain, but might create problems for integration in the TC211 HMMG. Therefore I would strongly recommend to consider that the UML description in this "Coverage Implementation Schema" be considered as not formally validated by OGC nor TC211.

As a result, in the context of INSPIRE, where specifications and schemas are supposed to be derived from standardized (and validated) UML models, I would urge the involved experts and stakeholders to double-check whether the INSPIRE Coverage model has an interest to upgrade to this new model (that is not available as an EA - or equivalent), and evaluate how the existing data specifications may be upgraded / mapped to the new schemas.

Hoping this helps

Material and documentation presented in the webinar

• Presentation:

• Excel spreadsheet - Description of Alternative 1 & Alternative 2:

• Alternative 2 applied to Elevation UML diagram:

• Alternative 2 applied to Orthoimagery UML diagram:

Webinar recording

• Streaming: https://ecwacs.webex.com/ecwacs/ldr.php?RCID=ce9335a4b7bfcd72c4e28851a2aad19c
• Download: https://ecwacs.webex.com/ecwacs/lsr.php?RCID=be28a5e39a4edbb3523f954eff66718b

Conclusions / Agreements

Two alternatives were presented in the webinar in order to solve the current issues identified by the cluster related to INSPIRE coverage data implementation:

• Alternative 1: Solutions based on the application of the OGC CIS v1.1 standard, recently approved.
• Alternative 2: Solutions based on the application of the OGC CIS v1.0 standard (namely GMLCOV).

It was a long webinar, but helpful to explain the audience the current issues to address and the possible ways to proceed (solve them), as well as to check with the cluster’s members which is the most appropriate alternative at the moment.

The general outcome was that Alternative 2 is the more pragmatic solution, at least while CIS v1.1 is not widely adopted by software packages publishing WCS services.