[gnso-rds-pdp-wg] Contactability

[Andrew Sullivan]

On Wed, Nov 29, 2017 at 10:20:33AM +0100, Volker Greimann wrote:
> Hi Andrew,
> 
> I guess our interpretations of the word "required" are different. It is
> certainly beneficial and helpful to have that system, but to say it is a
> conditio sine qua non is going to far as is aptly demonstrated by major
> ccTLDs not devolving into chaos despite removing the direct contact methods.

Yes, we are using the term differently.  My meaning of "required" is
"does design of the system depend on the candidate requirement?".
Yours is "does it work most of the time without that candidate
requirement?"

I agree these are both meaningful questions.  I agree that as an
empirical matter most of the DNS works most of the time without people
needing to contact one another (for if it didn't, we'd have ditched it
long ago instead of baking it into everything).  But I claim that my
meaning is a matter of accepting the logic of the system, and that the
empirical consequences of failing to embrace that logic do not provide
an argument about whether we are violating the logic of the system
design.

Internet systems are also designed to work remarkably well even in the
face of their assumptions being violated.  That's part of the genius
of best-efforts networking.  But in a best-efforts situation, it's a
shame to hobble those efforts.

Best regards,

A

-- 
Andrew Sullivan
ajs at anvilwalrusden.com