[Requests] public comment on candidate ‘GeoPackage Extension for Tiled Gridded Coverage Data’ standard

Keith Ryden kryden at esri.com
Tue Sep 19 18:51:07 EDT 2017



The following comments are submitted in response to the public comment period for the OGC GeoPackage Extension for Tiled Gridded Coverage Data

Page 6: The Applicability section seems both simultaneously too verbose and unspecific.  It does not match the Applicability sections for the other extensions – it reads more like a description.

Pages 13-14, 16: This is a general comment about extensions adding additional columns and applies to the base Geopackage spec (Section on extensions) as well.  Requirement 120 reads:

A GeoPackage that contains tiled gridded data SHALL contain a gpkg_2d_gridded_coverage_ancillary table or view as per the Coverage Ancillary Table Definition below.  Subsequent extensions or custom implementations MAY add additional columns to this table.  Clients SHALL ignore additional columns that are unrecognized.

The problem with this statement is that it leaves the door open for extensions to add columns that may have constraints that could be inadvertently violated by applications – thus making it potentially impossible to ignore additional columns that are unrecognized.  The GeoPackage Extension mechanism should explicitly state that any added columns should not have constraints that would prevent an insert/update/delete from an application that does not recognize the columns.

Page 17: The extension defines properties for min/max/mean/std-dev for every tile at every scale – thus values for a specific sample.  Not real clear how useful this is – but more importantly, how do you get an unambiguous min/max/mean/std-dev for the entire gridded coverage?

Page 19: Is it correct for both the metadata tables and any data tables to be listed in the gpkg_extensions table under entries for this extension. This is probably a clarification for the general Extension mechanism description in the base GeoPackage document – if an extension has metadata tables, should they be listed in the gpkg_extensions table?

Pages 23-24: Req/gpkg-tiff-internal-files ought to be Req/gpkg-internal-tiles (with a ‘t’ instead of an ‘f’).

General comment on requirement numbering – this document is separate from the GeoPackage base specification.  The requirement numbers should be specific to this document and start at 1.  As it stands, the requirement numbers appear to start at a random point (I suspect the next available requirement number in the base specification).


Regards,

Keith Ryden
Esri Software Development
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opengeospatial.org/pipermail/requests/attachments/20170919/f57694f5/attachment.html>


More information about the Requests mailing list