Esszett, Final Sigma, ZWJ and ZWNJ

[Cary Karp]

> I agree that making specific bundling recommendations is way outside
>our scope.  I do think, however, that a nearly-complete outline of how
>to solve the problem that underpins Paul Hoffman's objection would do
> an awful lot to address that objection.

I'm not certain if this addresses the concern to which you are
referring, but hope it may prove a useful contribution to the
discussion, in any case --

One way that a zone administrator could process requests for the
registration of a label including an Esszett would be to compare the
label in which it appears with all of the labels that are already
included in the zone. If a pre-existing label includes a double s
starting in the same position that the Esszett appears, and the two
strings are otherwise identical, if the new label is being requested by
the holder of the previous label, the registration will be accepted. If
the new label is being requested by a separate entity, the requester is
advised that a potentially conflicting label is already included in the
zone and that the registration will not necessarily be accepted. At the
same time, the holder of the pre-existing label is advised of the
registration request and asked to acknowledge it. If it is
acknowledged, the registration of the new label is accepted. If it is
not acknowledged, the earlier registrant is afforded opportunity to
register the new label. If that is utilized, the registration of the
new label is finalized. If it is rejected, the zone administrator can
assess the reasonability of blocking the registration entirely, for
example, if the two labels are clearly derived from German lexemes and
can reasonably be taken as orthographically equivalent. If there is
no such equivalence, a case to case decision will be made about
permitting the latter applicant to register the label containing the
Esszett.

Resulting disputes fall under the UDRP, which is facing an
IDN-conditioned nightmare, in any case.

On the protocol side, we could discuss the utility of establishing a
class of contextual rule that would  operate on separate labels. The
Esszett could then be permitted contingent upon the existence of such a
rule. Please note that I am not proposing this action -- just trying to
find a way to bridge the confidence gap that appears to exist between
the protocol developers and the operators of (primarily) TLDs.

/Cary