Re: [dispatch] New version for the Alert-Info URNs draft: draft-liess-dispatch-alert-info-urns-00.txt

[Adam Roach <[e-mail]>]

On 10/19/09 9:38 PM, [e-mail] wrote:
>     - Alert-Info URNs Indications for country-specific tones
>    

While I think this is a good dimension to have, I question the wisdom of 
replicating the entire ISO 3166-1 registry in an RFC and into IANA, for 
a variety of reasons:

   1. The initial registered list -- the one in the RFC -- will rapidly
      become out of date.

   2. The IANA registry will require constant updating to match the
      changes in ISO 3166-1. This either requires the IANA to keep up
      with changes in that registry (which is a task well beyond what we
      require of them at the moment) or it requires someone to send
      notes to IANA to update the registry whenever a change is made. We
      don't really have the structure necessary to ensure that someone
      will do that.

   3. If IANA and ISO 3166-1 do become out of sync -- which would almost
      necessarily happen from time to time -- you'll end up with
      implementations selecting at random between them (which is not a
      recipe for interoperability).

   4. The selection of country codes is highly political and rife with
      land mines -- some are forseeable (e.g. tw, fk), while others
      aren't necessarily. You can try to side-step this -- like ICANN
      tries -- by claiming that we don't take codes outside of ISO
      3166-1, and that we must accept all codes in ISO 3166-1; however,
      that hasn't kept them out of occasional hot water.


I would strongly suggest simply citing ISO 3166-1, not putting any 
values in the IANA registry, and simply giving a couple of 
representative (and politically uncontroversial) examples.

/a