Re: new ECML v 2 requirements from Nokia

[Eric Brunner <brunner@world.std.com>]

Donald,

> Whether you like the situation or not, the Palestinian Authority

We don't dork with ARIN/RIPE/APNIC addr allocations because some State
has a preference for negative-allocation to a competitor.
We don't dork with the IANA root because some State wants to edit it.
We don't dork with BGP4 because some State wants to prevent peering to
a competitor.

We are neutral, or as neutral as we are conscious of how mechanism
is distorted by policy. States don't define the scope of correct
function by a protocol, even an application protocol.

Some octets for an identifier. Is Big. Fine.
Two octets interpreted as 16 bits. Is Still Big. Fine.
Two octets restricted to ASCII isalpha(). Not Is Big. Worry.
Two octets restricted to ASCII isalpha() and restricted to fewer than
300 values, over a hundred of which are absolutely bogus for a 1st-world
financial infrastructure except as tax dodges and worse, and which makes
extension elsewhere manditory. Ouch.

> So you don't like iso3166. What exactly is your alternative? 

Don't overspecify. Use MAY nor MUST. It isn't our business to codify the
passions of Kings and Presidents, or make the net safe for Nation States.

I object to iso3166 as unnecessary overspecification of the range of
identifiers of a two- (or more) octet, octet-alligned field. You've
noted the objection, so we're probably done.

Eric