Is there even the slightest thought of producing an object oriented schema for geospatial information? The present structured (as in: structured analysis and design - not to be confused with composition) approach to schema forces a lot of extra coding on an OO system.

An OO system must either 1) parse the attributes of the inbound structure to ascertain which subclass to cast, or 2) parse the attributes of a generic (a.k.a. generalized) class to know which function to use when performing a computation. Hopefully the implementers use the first option vice the second, otherwise I'm paying the cost of using OO languages twice - once because we're using an OO language and twice because we're using structured code (which has to perform all the dispatching work).

Thanks for listening.

Very respectfully,



If you want a concrete example: UOM is a specialization of measurement in OO thinking - not an attribute. (Yes, there are several who will argue this because the calculation of area, for example, is the same no matter your unit of measure - but consider the circumstance where UOM is not the same on all information or precision varies across information.)

creed's picture


The OGC Geography Markup Language is grounded in a number of abstract models, such as ISO 19107 - Spatial Schema. This abstract model is documented in UML (in Enterprise Architect). As such, the model (or GML) can be mapped into an OO structure. For example, check out: http://portal.acm.org/citation.cfm?id=1183471.1183490 . As to the current work in the OGC, there are no near term plans to define an OO encoding.