XRI/XDI - Xref Delimiter proposal

In the past few weeks, a syntactic construct not currently valid under XRI
2.0 has been repeatedly mentioned in different XRI 3.0 scenarios: treating a
cross-reference (parentheticals) as a standalone subsegment when not
immediately proceeded by another XRI delimiter (when proceeded by another
delimiter, either GCS or LCS character, the cross-reference is part of that
subsegment).

This came up independently in XDI TC discussions of reification; in
discussion of the XRI 3.0 resolution function at the Friday F2F; and again
today in Nick's message and my discussion with him tonight.

I have formalized this as the "Xref Delimiter" proposal at:

	http://wiki.oasis-open.org/xri/XriThree/XrefDelimiter

Alongside the GCS Delimiter proposal
(http://wiki.oasis-open.org/xri/XriThree/GcsDelimiter), this would now mean
all constructs in XRI 3.0 syntax would be "atomic", i.e., by themselves they
constitute an XRI subsegment unless preceeded by an XRI delimiter of higher
precedence. The precedence order is:

1) Global context symbol (=, @, +, $)
2) Local context symbol (*, !)
3) Cross-reference (parens)

See http://wiki.oasis-open.org/xri/XriThree/XrefDelimiter for examples.

I'll put this on the agenda for this Thursday's telecon.

=Drummond