Composite Name and Naming System Boundaries

There might not be a one-to-one correspondence between component separators and naming system boundaries if a composite name contains names from naming systems that use the same character as the XFN component separator to separate their atomic names. Consequently, a component of a composite name might represent an atomic name from a hierarchical naming system that uses the XFN component separator or a compound name. Strong separation and weak separation refer to how a context considers the XFN component separator as a naming system boundary.

Strong Separation

An XFN context that treats the XFN component separator as a naming system boundary supports strong separation. An XFN component separator that appears within a component to be resolved by the context must be escaped or quoted.

Support for strong separation is a property of a context. A context that supports strong separation expects to receive the name that it is going to resolve entirely in one component of the composite name structure. When a composite name is supplied to such a context, it consumes the leading component of the name; any remaining components are left to be resolved by subordinate naming systems.

An XFN context with a name syntax that is either flat or hierarchical, and does not use the XFN component separator as its atomic separator, supports strong separation. Examples of naming systems that support strong separation are DNS and NIS+, both of which have right-to-left dot-separated names. The following are examples of names with DNS and NIS+ components, respectively.

	.../wiz.com/orgunit/ppt
	orgunit/accountspayable.finance/user/jsmith

Weak Separation

An XFN context that does not always treat the XFN component separator as a naming system boundary supports weak separation. This arises when the component naming system associated with the context uses the same character as the XFN component separator as its atomic component separator. The context allows its atomic separator to appear unescaped and unquoted in its compound names when they occur in composite names. This means that an XFN component separator might not necessarily signify a naming system boundary.

Support for weak separation is a property of a context. A context that supports weak separation expects to receive its atomic names in separate components of the composite name structure. When a composite name is supplied to a context that supports weak separation, the context consumes the leading components of the name (and treats them as atomic components); any remaining components are resolved by subordinate naming systems. The number of components consumed is determined either syntactically or dynamically.

CDS names and X.500 names are examples of names that use the XFN component separator as their atomic name separator. X.500 supports weak separation using a syntactic method (by scanning for typed names) while CDS supports weak separation by determining the naming system boundary dynamically.

The following example shows a composite name with an X.500 component.

	.../c=us/o=wiz.com/orgunit/ppt

An XFN context that supports weak separation using only syntax-specific discovery of its naming system boundary might not always be federated with arbitrary subordinate naming systems. If the subordinate naming system has a naming syntax that is indistinguishable from that of the superior naming system, the superior naming system is not able to identify the naming system boundary.

Naming systems that use the same character as the XFN component separator as their atomic separator, and which cannot support weak separation because it cannot use a syntactic or dynamic method to determine the naming system boundary, must provide context implementations that support strong separation. This means that occurrences of atomic separators must be quoted or escaped when they appear in compound names within composite names.