1. Create a GML application profile schema based on LandXML-1.0, coined "LandGML".

2. LandGML enables GML 3.0 aware applications to read the geometric data without requiring software code changes or any special knowledge of LandXML.

Goal #2 was the real reason behind the production of the two software translation tools as opposed to Geospatial applications reading LandXML files directly. Although at the time of this writing there are commercial Geospatial applications available that import LandXML directly.

Results:

Goal #1 Success - Both transformation software tools (LandXML->LandGML & LandGML->LandXML) were built and proved bi-directional data transfer was possible.

Goal #2 Failed - In the summer of 2004, there was only 1 GML 3.x aware "alpha" application to test with and it proved that software code modification was indeed required to support the data.

While proving the academic data modeling point for goal #1, no commercial or practical viability presented itself. Commercial viability is the real measure of any data standards success.

An important lesson learned is the fact that most LandXML-1.0 files do not contain a specific coordinate system, which makes inserting the data into Geospatial applications problematic. A proposed fix for this is proposed in the new LandXML-1.1 schema specification.

Final Comments:

The LandGML schema was based on 26 sub-schemas that comprise "GML" and proved to be nearly 4x the size of the LandXML schema file. While file size for the schema file is not a big deal, digging into and trying to understand each element and its derivation hierarchy requires a traversal of 1 to 26 sub-schema files, which is a very big deal if you are trying to write software to support it.

LandXML instance documents converted to LandGML instance documents initially proved to be 20x larger than the original LandXML file size, but was later refined to only 2x-4x larger. What was gained by creating a more complex schema and much larger instance files? Nothing - but what about "off-the-shelf" GML applications that can read the data? At the time of this writing there aren't any GML 3.x compliant applications commercially available. The OGC website indicates 16 applications are in the "implementing" stage and have been in the same status for more than a year. Once several commercial GML 3.x applications are available this experiment should be revisited.