Use case: CRS specification

Background

The discussion was triggered by a question in Frans Knibbe's mail of 20 Dec 2013, about the possibility of specifying a CRS separately from the geometry representation / serialisation.

A specific thread was opened on this topic on 22 Dec 2013, and continued in Jan 2014. However, the discussion was carried out in parallel also on the original thread.

Summary of discussion

The possibility of specifying a CRS separately from the geometry representation / serialisation was supported by two main use cases:

1. Ability to explicitly specify a CRS in data aggregated from different sources, which are using, implicitly, different default CRSs - see Raphaël's mail (22 Dec 2013)
2. Ability to query the CRS (e.g., filter geometries based on their CRS) also in SPARQL endpoints not supporting GeoSPARQL or stSPARQL - see Frans's mail (30 Dec 2013)

The discussion basically focussed on two main issues:

• The possible drawbacks of specifying the CRS separately from the geometry representation.
• Whether it would be possible or not to address the use cases with CRS specified as part of the geometry representation / serialisation.

The group moved towards the agreement to allow both approaches as alternative and not exclusive options for the specification of a CRS. It was also considered the possibility to add an optional property to the LOCN vocabulary for specifying the CRS separately from the geometry representation / serialisation.

About the domain of such property, in his mail, Frans argued that the CRS should be applied also to collections of geometries and datasets. No final decision has been taken on this issue.

Finally, another discussion topic was about how the CRS should be denoted - by a literal, URN or URI. There was a general agreement that CRSs should be specified by using HTTP URIs, but the question on which registry should be used was left open.

Email reference

The relevant emails (in chronological order):

• mail from Frans Knibbe, 20 Dec 2013
• mail from Ghislain Atemezing, 20 Dec 2013
• mail from Oscar Corcho, 22 Dec 2013
• mail from Raphaël Troncy, 22 Dec 2013
• mail from Oscar Corcho, 22 Dec 2013
• mail from Andrea Perego, 22 Dec 2013
• mail from Raphaël Troncy, 22 Dec 2013
• mail from Krzysztof Janowicz, 22 Dec 2013
• mail from Raphaël Troncy, 27 Dec 2013
• mail from Raj Singh, 27 Dec 2013
• mail from Ghislain Atemezing, 27 Dec 2013
• mail from Raphaël Troncy, 27 Dec 2013
• mail from Andrea Perego, 29 Dec 2013
• mail from Raj Singh, 30 Dec 2013
• mail from Frans Knibbe, 30 Dec 2013
• mail from Clemens Portele, 30 Dec 2013
• mail from Frans Knibbe, 2 Jan 2014
• mail from Kostis Kyzirakos, 2 Jan 2014
• mail from Ghislain Atemezing, 3 Jan 2014
• mail from Raphaël Troncy, 3 Jan 2014
• mail from Raphaël Troncy, 3 Jan 2014
• mail from Frans Knibbe, 3 Jan 2014
• mail from Kostis Kyzirakos, 3 Jan 2014
• mail from Kostis Kyzirakos, 3 Jan 2014
• mail from Raphaël Troncy, 3 Jan 2014
• mail from Frans Knibbe, 3 Jan 2014
• mail from Raphaël Troncy, 3 Jan 2014
• mail from Simon Cox, 6 Jan 2014
• mail from Simon Cox, 6 Jan 2014
• mail from Raphaël Troncy, 6 Jan 2014
• mail from Kostis Kyzirakos, 6 Jan 2014
• mail from Kostis Kyzirakos, 6 Jan 2014
• mail from Frans Knibbe, 6 Jan 2014
• mail from Raphaël Troncy, 6 Jan 2014
• mail from Raj Singh, 6 Jan 2014
• mail from Ghislain Atemezing, 6 Jan 2014
• mail from Kostis Kyzirakos, 6 Jan 2014
• mail from Kostis Kyzirakos, 6 Jan 2014
• mail from Raphaël Troncy, 7 Jan 2014
• mail from Andrea Perego, 9 Jan 2014
• mail from Andrea Perego, 13 Jan 2014
• mail from Oscar Corcho, 13 Jan 2014
• mail from Frans Knibbe, 14 Jan 2014
• mail from Bart van Leeuwen, 14 Jan 2014
• mail from Kostis Kyzirakos, 14 Jan 2014
• mail from Raphaël Troncy, 18 Jan 2014