Stability of valid IDN labels

[Eric Brunner-Williams]

Mark,

The requirements for labels arises somewhere, and I suspect that for the 
IETF's IDN activities, past and present, the author of that requirement 
is, at in part, ICANN, in the broad, constituency sense, not the almost 
uniformly non-technical members of staff employed by ICANN.

If that is the origin of the requirement for stability, in my opinion, 
your requirement that a label be "valid (or invalid) forever" is 
stronger than the requirement originating from ICANN for IDNs.

Please consider the 3166 code points in the zone file that the IANA 
directly manages. These are not "valid (or invalid) forever", where 
there was an alpha-two label "SU" there are now several distinct 
alpha-two labels labels. The ISO3166/MA publishes newsletters with 
occasional changes to one or more code points, which are eventually 
reflected in the validity, or invalidity, of labels in the root zone. 
Now it is true that there is no change in the [A-Z][a-z][0-9][-,.] and 
length restrictions on labels formed from ASCII characters, but that is 
a truth about a production rule, not a truth about any particular label.

Perhaps this is a distinction without difference, as per-character 
determinations of label validity may be a production rule for a class of 
labels with only one member.

If so, again in my opinion, the ICANN requirement for stability does not 
approach the semantics of "forever", there simply is no planning process 
in ICANN for any requirement that is as fixed in its intent as that of 
the authors of the BCD series of standards, or the ASCII series of 
standards, for a present work that cannot, and also need not, be 
modified within a few short years, as we are attempting now, on the work 
of the earlier IETF IDN WG.

The difference between "for as long as we think this work is used" and 
"forever" may allow CONTEXT rules to become more restrictive.

Eric