Re: [idn] Intro to my I-D

[Eric Brunner-Williams in Portland Maine <brunner@nic-naa.net>]

Ben,

Assume Simplified characters X, Y and Z, each have just one
Traditional form, characters U, V and W, respectively.

The names formed from the union of these three equivalence classes
are:
	XYZ, XYW, XVZ, XVW, UYZ, UYW, UVZ, UVW

If the registrant is able to do the following as he or she so chooses:
XYZ.cn (ip addr 1.2.3.1)
XYW.cn (ip addr 1.2.3.2)
UYZ.cn (ip addr 1.2.3.3)
UYW.cn (ip addr 1.2.3.4)
XVZ.cn (ip addr 1.2.3.5)
XVW.cn (ip addr 1.2.3.6)
UVZ.cn (ip addr 1.2.3.7)
UVW.cn (ip addr 1.2.3.8)

Then the following is possible, neh? (Assume subnet of 1.2.3 throughout)

+-------+-------+-------+-------+-------+-------+-------+-------+---------
Number of characters with only one equivalent character (SC/TC, simple case)
1 (2)		2 (4)		3 (8)		4 (16)		...
Registrat A	Registrant B	Registrant C	Registrant D	...
+-------+-------+-------+-------+-------+-------+-------+-------+---------
				XYZ.cn (1)
				XYW.cn (2)
                XY.cn (127)	UYZ.zn (3)
X.cn (254)	UY.cn (127)	UYW.cn (4)
U.cn (253)      XV.cn (127)	XVZ.cn (5)
		UV.cn (127)	XVW.cn (6)
				UVZ.cn (7)
				UVW.cn (8)
+-------+-------+-------+-------+-------+-------+-------+-------+---------
multiple	unique		multiple	...		...

I'm glad you intend to comment on any draft, whether by writing a comment,
or an alternative draft. However, this appears to require users to have
access into the ideosyncratic (private) semantic preferenances of the
registrant. How, as a set of labels is expanded from one initial character
to some maximum number of characters, each a single character suffix to
the preceeding label's set of characters, e.g., a, ab, abc, abcd, ...
are labels to be understood? As having consistent equivalence rules, or as
having label-unique equivalence rules? All this is within a single zone.

Not that I consider the interests of trademark holders to be the only ones
worth examining, but if some set of SC and TC characters are registered
marks in some jurisdiction, then how would a mix of SC and TC characters
from both registered marks appear to the trademark holder? It appears that
the advantage of your proposal is that users must engage in vastly more
complex name-to-address management, as must marks enforcers.

Your initial note claimed the benefits are mapping content script to label
script, and user-defined, reserved character equivalence definition. Have
I got your proposal's main points?

Eric