[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: New Internet Draft on registering IDNs



Martin & all,
    
Your next questions/examples was something I was thinking about
some time.  We (can) have similar problems with Dutch as well, but
I keep my comments to your .de example.

    For example, an u-umlaut in German is often equated with 'ue'.
    Now suppose that the German TLD wants to define tables to take
    this into account, and apply these to the current ASCII registrations
    in .de. Can they design the tables so that they don't have to take
    away already registered domains from their owners? If they apply
    the algorithm, in what sequence do they apply it to the registrations
    they already have? Or, how do they design the tables so that there
    is no difference in the result independent of what sequence they
    apply it? These are all specific, real, questions.

Yes, these are real problems. There is the domain name clueless.de
(and clue.de and probably more such as feuhrer.de). If the algorithm
specifies that u-umlaut and ue are in a single bundle it will might
mean that either the clue.de delegation should be cancelled or that
nonody can register clu+umlaut. That last one is probably unlikely,
but feuhrer.de and fu(+umlaut)hrer.de (or any other word that makes
sense and is already registered)  is a potential clash. There are
probably more of these clashes then I can think of, but I leave to
more quelified German speakers then me to come up with them.

    If there are easy answers to them, it doesn't hurt to state
    them in the draft.  If these are unsolved questions, then we
    can state it so that we can start trying to solve them.

The easy answer is of course that denic decides not to introduce
the u+umlaut at all and just live with the current status quo. But
that is quite unsatisfactory.  I'm afraid I don't see even a start
to solve this type of problems.

	jaap