Hi Nick, George

"fundamentally yeah - it's a bit colonial to continue ignoring the 
reality that us-ascii is inappropriate for large chunks of the world."

I totally agree with the drive to disengage with the colonial past and move on in all areas of life. However, the RIR Databases reflect a public global registry of IP addresses, and in parts also serve as IRRs. If we are going to allow all character sets within these databases then we must answer a number of non technical questions, such as:
-What is a global IP address registry?
-What purpose does it/do they serve?
-Who is it for?
-Who needs to be able to read, understand, interpret, convert what parts of it?

Before we even ask those questions we need to consider who is going to answer these questions and on whose behalf? Is this going to be addressed separately by each RIR? Or is it better to consider a cross RIR Database/Registry solution?

This is why I believe we should go for a simple Punycode option now and then start to consider the bigger picture and finally address difficult questions that have been brushed under the carpet for many years.

cheers
denis

co-chair DB-WG


On Monday, 13 July 2020, 16:42:01 CEST, Nick Hilliard via db-wg <db-wg@ripe.net> wrote:


George Michaelson via db-wg wrote on 10/07/2020 01:15:
> Raw UTF-8 would work better here. It would permit the model to be
> extended naturally into multiple script models. I understand its
> inherently more complex than uplift to punycode of the domain elements
> in things, but the underlying problem in Whois and language goes
> beyond the specific domain-name context.
>
> I guess I am saying "I agree with Gert and I think my community wants this too"

fundamentally yeah - it's a bit colonial to continue ignoring the
reality that us-ascii is inappropriate for large chunks of the world.


Nick