There are label segments that have semantics.
The "--" violation, prepended by something, "^xn", where "^" indicates
a label boundary, to indicate a (the current) "IDN" processing.
Bytes within a label with values in excess of 127.
Off hand I can't think of anything else (that is intentional). Along
the accidental sort-of-intended mumble access of forensic engineering
are terminating (TLD) labels that brain-dead dead-ware interprets (if
anyone living were to ask it) character representations of digit
sequences as v4 addresses that currently has the pants scared off of
ICANN and some folks from this side of the policy|tech divide.
Everything else seems to just be intentional carve outs of string
spaces, suited, if there is a purpose, to the (defunct) User Services
Area Directorate and its inform-the-community mission that "example"
is not a means to repeat for the bazillionth time the flagship brand
of Verisign.
.local has sufficient meaning to care about.
John Levin's draft covers some "don't pretend you're ARPA" ground,
which is a useful restriction on what appears in the IANA root (and
any other used by every user in China) and in SLDs, assuming that a
means other than persuasion is available to inform both
auctions-are-good gTLD operators and ICANN-is-irrelevant ccTLD
operators that some restrictions are beneficial.
The chief defect of Stewart's draft is that it makes an analogy to the
semantics of addressing, and postulates a pseudo-technical set of
implementation responsibilities, and fails to mention ICANN, which has
some coordination mission.
What isn't a layer 9 is worth documenting. What is at layer 9 should
be documented as being at layer 9.
Eric
_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf