Stability of valid IDN labels

[John C Klensin]

--On Tuesday, 22 April, 2008 19:43 +0200 Frank Ellermann
<hmdmhdfmhdjmzdtjmzdtzktdkztdjz at gmail.com> wrote:

> John C Klensin wrote:
> 
>> To say this a little differently, such changes in Unicode are
>> not made in secret, but involve opportunities for public
>> discussion.  It is reasonable to assume that the negative
>> effects on IDNs of making such changes would be part of the
>> relevant discussions.
> 
> I'm far from sure how reasonable that is:  The introduction of
> u+1E9E was discussed in public, but decided by a third party
> not interested in its effects on the Internet, let alone IDN.

Frank, enough people who do care about IDNs watch the Unicode
list, and several members of UTC have been engaged enough in
these discussions, that I believe IDN impact, and consequences
to other Unicode applications that are very sensitive to
stability (very strongly defined), will be raised if property
changes are contemplated.  I do not assume that those
discussions will always resolve themselves in a way that is
favorable to the needs of IDNs: there may be other
considerations that the Unicode Consortium concludes outweigh
the IDN issues.  But the latter are precisely why that "backward
compatibility" category exists in the Tables document.  It is
not critical for us that such changes never be made, only that
they do not catch us completely by surprise.

Of course, should the Unicode Consortium start behaving in a way
that was completely irresponsible (as distinct from making an
occasional decision which we might not like from an IDN
standpoint), we would have far more serious problems than we can
anticipate and address in these documents.  But I don't see any
reason to predict that behavior or justification for spending
time planning for the possibility.  And the more Unicode is used
for true multi-language and multi-script applications, rather
than as a way to deal with individual scripts while avoiding
state-switching, the more pressures from the entire user
community will act against any excesses.  One of the
disadvantages of having a successful standard is that it gets
really hard to change anything significant, even if one
concludes that would be a good idea (look at Internet email).

    john