"Seegrid will be due for a migration to confluence on the 1st of August. Any update on or after the 1st of August will NOT be migrated"

Discussion: Geologic Relation

Also see GeologicRelation

Tucson Meeting Notes

GeologicRelationBruceSimons (GeologicRelation and GeologicRelationDiscussion) * Topics * Is it needed? * Should it be a Feature Type or Object Type?

  • Discussion * need something to be able to document a fault cutting a unit or one unit overlying another unit * If specific would need to add it to unit to unit, structure to unit and structure to structure * two ways of doing it * leave as is and use generic geologicalFeature relation for all cases, but would need to remove some specific cases already in the model

role attribute

role is defined as:"Term that specifies the relationship between source and target geologic features." This is ambiguous. Is it specifying the relationship type (eg 'intrusive', 'stratigraphic', 'spatial') or the role each particpant is playing in the relationship (eg 'intrudes'-'hosts'; 'overlies'-'underlies')?

If the former, how is each participants role described? If the latter where is the relationship type specified and is it the target role or source role?

-- BruceSimons - 29 Sep 2006


Should GeologicRelation be a FeatureType? This would allow relations to be served by WFS.

Need to clarify top-level hierarchy. What are GML Objects?

My current thinking is that there is gml:AbstractGMLType, and then should only be the three direct specializations that already appear in GML, viz:
  • gml:AbstractGMLType is the abstract root for things with identity
    • since this is abstract, there is no service interface
  • gml:AbstractGeometryType is the root for types whose instances are geometry objects, whose only properties are related to shape
    • there is currently no direct service interface but there probably should be one
  • gml:DefinitionType is the root for types whose instances are concepts, whose definitions are relatively static and under formal governance arrangements,
    • usually served by WRS (Catalogue), but are stable and may be cached
  • gml:AbstractFeatureType is the root for types whose instances are feature instances, from a source whose update is controlled by a service instance provider under their own rules
    • collections are served by WFS, and may change between requests

Everything in an application schema is thus either
  1. a geometry
  2. a definition
  3. a feature

Under this analysis, Relations are features.

This is also supported by the "definition" of features in ISO 19109 ("Rules for Application Schema") clause *7.3.2 The purpose of the GFM". For example:
... the concepts in the GFM establish a basis for the classification of features, ...

The things we want to classify we call features; the relations between feature types are feature association types and inheritance. Feature types have properties that are feature attributes, feature operations and feature association roles. ...

-- SimonCox - 02/06 Jun 2006

I'm not sure I understand your reasoning. gml:AbstractGMLType is abstract but classes derived from it are not. Furthermore, my feeling is that WFS is broken (it was really thought out GML 2 / GML 3 Level 0). Changing the model just to have it compliant to WFS is not a good argument in my opinion.

But having Relations to be features makes sense. Relations have identity, a location and a timespan, they seems to fit the definition of features, I think.. maybe..

this is my opinion after my second coffee on Sunday morning-- not that it could not change.

-- EricBoisvert - 04 Jun 2006


RelationElement, change datatype on GeologicRole to CategoryValue, change data type on proportion to QuantityValue.

-- SteveRichard Date?

Do these attributes still exist in the current model? I couldn't find them.

-- BruceSimons - 29 Sep 2006

See earlier discussion in CgiRelation


  • Why is GeologicRelation a FeatureType ?
  • Do we need GeologicRelation in GeoSciML ?

-- BoyanBrodaric, EricBoisvert, BruceJohnson - 11 Apr 2007
Topic revision: r9 - 15 Oct 2010, UnknownUser

Current license: All material on this collaboration platform is licensed under a Creative Commons Attribution 3.0 Australia Licence (CC BY 3.0).