Hi Ben:

> I am not so sure. The GF_PropertyType in ISO 19109 is an abstract  
> supertype of GF_AttributeType. I think this means that, if GeoAPI  
> represents the ISO Feature Model, a Feature with complex properties  
> should have them as immediate properties, not enclosed in wrapper  
> complex properties like gml:FeaturePropertyType.

That is required in order to allow for associations. We do not have a  
separate accessor for attributes and associations - they are returned  
as a single list of properties.

>> Can you be more specific - I am not sure what I am looking at in  
>> these  UML models.
>
> In the UML for GeologicUnit, you can see that a GeologicUnit has  
> 0..* CompositionPart. Apologies for my ASCII UML:
>
> GeologicUnit <>--- 1..* CompositionPart
>
> In GML (the GeoSciML application schema), a GeologicUnitType has  
> 0..* CompositionPartPropertyType properties, each of which contains  
> a CompositionPartType:
> http://www.geosciml.org/geosciml/2.0/xsd/geologicUnit.xsd
>
> In UML, the XML looks like:
>
> GeologicUnit <>--- 1..* CompositionPartProperty <>--- 1  
> CompositionPart
>
> and the encoded XML has XPaths like this:
>
> gsml:GeologicUnit/gsml:composition/gsml:CompositionPart
>
> It seems to me that in GML an enclosing type has been introduced as  
> an artifact to ensure that every complex property contains an  
> element that gives its type information. To me this seems to be an  
> encoding artifact that should be created at encoding time, not  
> stored in our GeoAPI representation of the data.

That does not seem too strange.  In this case I am guessing that  
GeologicUnit is a bit dynamic in nature; can probably form a hierarchy  
so they are needing to provide type information as they go to account  
for the dynamic nature of their problem?

It is unclear to me if this is a problem to be fixed; or a comment on  
how interesting the GeologicUnity is? Perhaps I am missing something?

> GML is only a representation of the GeoAPI model.

Yep.

> In short, should GeoAPI represent the General Feature Model or the  
> GML encoding? Or am I missing something?

The "General Feature Model" is kind of defined by a couple  
specifications in both ISO and OGC land. GML has softened the general  
feature model a small bit; but general practice of GML has softened it  
even more. Our task is to bring things back the other way.

Can we refine what we have here based real world use - ie your  
experience here and now.

Jody

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
Geotools-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geotools-devel

Reply via email to