It is propably a good idea to decouple the definitions of TM Applications
from the definitions of interchange syntaxes.
Suggestion: Add a whole new top level section named "Definitions of
interchange syntaxes" addressing the requirements for such definitions.
Among these requirements must be the definition of a processing model
(propably change the term to 'deserialization specification') that
specifies how a given syntax is to be interpreted in terms of a
certain model.
Syntaxes are not bound to a ceratin model, only processisng models are.
A given syntax can be interperted in different ways by different TM
models.