Re: [idn] pool deadline imminent

[Keith Moore <moore@cs.utk.edu>]

> > > Don't you worry about replicating the Coloured Book insanity?
> >There have been so many different sets of coloured books (with different
> >kinds of insanity) that I don't know which one you mean.
> 
> Sorry.  Showing my age.  UK Coloured Book naming system that reversed the
> labels (at least according to 1034/1035 specification).  Despite programs
> and gateways designed to reverse the UK names, they leaked and resulted in
> lots of annoyance and confusion.

ah yes.  (oddly enough, that particular kind of colored book hadn't occured
to me, even though I'd written code to deal with it in a mailing list manager)

no matter what solution is chosen, we'll have leakage (in both directions) 
between legacy environments (that in general can only deal with ASCII names)
and environments that are explicitly written to deal with IDNs.   the major 
difference between the proposal camps is in how often the names leak, and 
what breaks when the names leak.  

> I am worried that a similar thing will happen with folks using ACE --
> people will (say) cut and paste a (utf-8 representation of a) name from an
> application into something that is supposed to do lookups (based on ACE)
> and those lookups will fail because a conversion from UTF-8 to ACE didn't
> happen.  To the user, this won't make sense since could look identical.

we can't avoid that case no matter what solution.  most legacy apps won't
be able to deal correctly with IDNs that are input in native format, period.

Keith