Proprietary schemas are by definition different in notation and description. When we build our own schemas there is no standard or fixed set of markup tags that define a common structure that is understood and shared between seperate departments or organizations.

To enable crosswalks for proprietary metadata requires that we provide this commonly understood and consistent structure.

JUMP does this by applying a higher-layer of semantic markup to a set of shared semantic objects.

<<semantic:1>> author <</semantic:1>>	<<semantic:1>> author <</semantic:1>>
<<presentation:1>> <book_author> <</presentation:1>>	<<presentation:1>> <creator> <</presentation:1>>
<<semantic:2>> publisher <</semantic:2>>	<<semantic:2>> publisher <</semantic:2>>
<<presentation:2>> <pub-co> <</presentation:2>>	<<presentation:2>> <publisher> <</presentation:2>>
<<semantic:3>> publish_date <</semantic:3>>	<<semantic:3>> publish_date <</semantic:3>>
<<presentation:3>> <pubdate> <</presentation:3>>	<<presentation:3>> <date> <</presentation:3>>

Here we apply the semantic markup and semantic objects as an overlay to the XML markup in each schema. This overlay structure makes it easy to recognize the correspondence between the two <<presentation:1>> tags, and between the two <<presentation:2>> tags, and so on. By simply aligning the shared semantic objects we can easily identify equivalent metadata between two different schemas.

This simple layer provides the necessary structure to allow us to build crosswalks between two different proprietary schemas.