RFC 5242 and IDNA

[JFC Morfin]

Dear John,
Got your last mail, futher I thought I closed the issue.  Let me 
comment on this to support your position and hopefully help everyone.

At 15:32 09/07/2008, John C Klensin wrote:
>I, at least, have been disappointed that some of the more subtle 
>parts of the joke have apparently not even been noticed, but so it goes.

I have been too. I tried to understand why because I think it is 
important to this WG, for the reasons you give, but also more fundamentally.

My reading is that you wanted to joke at what you want to understand 
from me, probably to show there was worse than what you had disputed. 
To do that you mixed an accurate description of complexity with 
deliberately foolish propositions. This leads to various ways of reading it.

People wanting to have fun at propositions they do not understand are 
not worried by another proposition they do not need to understand and 
enjoy the joke. People considering the proposition is a joke on a 
serious matter, consider what is serious under the joke. People 
informed of the issue, also consider what is serious (and there may 
be lucky good points, as in every joke)  but are dismayed by the lack 
of ingeniosity of the satire.

>I do suggest that, if the satire has any point other than humor,
>it is to illustrate that the design of a character coding system
>that uses completely consistent and unambiguous rules and that
>works easily and well for comparison purposes is devilishly
>difficult, if it is even possible (which I doubt).

This is THE key point. This is my main and only point: a single 
exclusive IDNA expected too much from Unicode.

>That
>demonstration is ultimately supportive of Unicode.  I have said
>many times (in different ways and in more serious contexts)
>that, while it is possible to find characteristics of Unicode to
>which to object and individual decisions to debate at length,
>the matter ultimately comes down to tradeoffs.

Yes!  The problem, IMHO is to want a single ballance of complex 
trade-offs for the entire world. This was the case in RFC 4646, it 
was the case  here with IDNA, until the clarification about ML-DNS. 
This is a layer violation between norms and standards, principles and 
applications. A very common problem - they grow but forget to 
multiply. A conflict of RFC 1958 recipes. Advantages of unilateral 
vs. power of diversity. This is the whole Tunis consensus deal - 
transition from Legacy Internet by WSIS non participating IETF to 
Emergent Internet by IGF - by the mechanism of enhanced cooperations, 
possibly unlocked by ICANN's current provocative TLD positions.

>There is
>considerable accumulated evidence by now that, while a different
>set of tradeoff decisions would yield a different result, there
>is no reason to believe that such decisions would yield a
>_better_ overall result than Unicode, only a different one.

Correct. This is why the objection is not to Unicode but to the way 
the IETF uses Unicode. IDNA is a trade-off that targets the Legacy 
Internet. ML-DNS will be a trade-off by the users of the emergent 
Internet. There may be others.

I fully agree with John. RFC 5242 is of no big interest. But what 
smart people may have learned from it should help everyone.

Now, may be could we resume the work on IDNA. My question about 
single Unicode TLDs and IDNs stands: is there a problem with them or not ?
jfc