Not sure if anyone's looked at Google's "LondonHouse" example for a Collada file exported from Sketchup, at
http://earth.google.com/kml/kml_21tutorial.html#models , but I assume it is representative of "good" Collada.
but I was wondering if anyone might comment on some interesting aspects of the schema....
In particular, it starts out with:
<COLLADA xmlns="http://www.collada.org/2005/11/COLLADASchema" version="1.4.0">
versus what is posted at collada.org,
<COLLADA xmlns="http://www.collada.org/2005/COLLADASchema" version="1.4.0">
which doesn't have the "11" .
Also, looking at the collada 1.4 .xsd schema doc, I see: <COLLADA> items can have assets, libraries, and scenes under it, that library is a complex type, which in turn can have: geometry, camera, controllers, etc under it.
So, parsers would expect: <COLLADA><library><geometry>...</geometry><...> </library> </COLLADA>
instead, I see <library_geometry>, etc type of tags hanging right off the root document, as indeed is described in the .pdf documentation.
So, my question is: should the .xsd be changed to reflect this reality? When I generate wads of classes purely based on the .xsd, the code generators take the schema doc quite seriously, and look for such structure.