Re: [idn] Adding "optional" characters in draft-ietf-idn-nameprep

[James Seng <James@Seng.cc>]

Paul Hoffman / IMC wrote:
> Whoops, I didn't understand you were suggesting that. I *strongly*
> disagree with the notion that an authority gets to have different
> matching rules. It will confuse the hell out of a user who, for
> example, gets one interpretation for foo.gr.com and a different one
> for foo.co.gr and a different one for foo.org.gr. And, as you state
> later, this messes with DNSSEC; we are required not to do that.

Actually, not neccessary per zone matching rule but more like per domain
matching rule. that would solve quite a lot of problem, including
allowing the registrant to define what kind of language rules they wish
to use for their domain.

For example, I send out the following proposal few months back...

--- CUT HERE ---
zone "foo.com" in {
    type primary;
    file "db.foo.com";
    variant "FOO.com" "FOo.com" "FoO.com" "fOO.com" "fOo.com" "foO.com";
}

[named will then fold all variant into foo.com automatically]

And on the parent zone, in this case, .com zonefile, you do

foo    NS     some.name.server.
FOO    DNAME  foo
FOo    DNAME  foo
FoO    DNAME  foo
>...

Is it nice? No. In fact, hell for operations. (I want to see the face of
the
people running large zone :-) 

But heck, it works. 

It can also be optimized if they can reduce it to some algorithm, which
make
life less painful for memory usage, by taking up more cpu cycle, 

For example,

foo    NS    some.name.server.
       FOLD  { s/I/*/g; tr/[A-Z]/[a-z]/; # * = dotless i }
; In this case, FOLD is not a RR, but some algo use for matching

Therefore, essentially, we can have "per domain" folding algorthim. 

Or perhaps more simple,

foo    NS    some.name.server.
       FOLD  tolower+utr21.

bar    NS    some.other.server.
       FOLD  tolower+simplifiedtraditional.

ie, with pre-defined algo.