RE: Who does what (Re: [idn] San Diego Meeting Notes)

[Patrik Fältström <paf@cisco.com>]

At 14.34 -0500 01-01-25, Brian W. Spolarich wrote:
>The authors do not state in IDNA why they believe IDNA
>is a better approach than IDNRA, which would be a helpful clarification of
>the draft.

My take is that it is better to do this in the application as the 
domainname is not only used in DNS but also in application protocols 
like SMTP. If we "only" fix the DNS protocol, and applications 
interaction with DNS, then we have yet another large problem to 
solve, and that is how to encode domainnames which are used as 
protocol parameters in each of the protocols using it.

By doing nameprep and encoding in the application, the same "output" 
of this transformation can be used in all communication on the net.

Also, if an application is not updated, regardless of how ugly 
leakage is, the user _can_ use the encoded string directly in the 
applications which today use as protocol parameter domainnames. I.e. 
I personally see this as an important feature just because 
replacement of software will not be as fast as some people might 
think.

I can add some more text on why it is better to have transformation 
in the application, but not a comparison between the two proposals. 
They are just I-D's and when one of them (or none) is published, the 
other (or both) will go away, and the reference to the non-existing 
document is no longer making sense.

    Patrik