[gnso-rds-pdp-wg] Proposed Agreement for Original Registration Date

[Andrew Sullivan]

On Thu, Sep 21, 2017 at 04:54:59PM +0200, Volker Greimann wrote:
> My issue with the counter is that it is just as prone to errors than the
> originally proposed date.

How?  The counter is "known previous registrations", not the "original
date", and it can be NULL.

For instance, suppose we have this timeline:

    1.  example.com exists with ROID 1234
    2.  example.com is deleted
    3.  example.com is created with ROID 2345
    4.  example.com is deleted
    5.  example.com is created with ROID 3456

Under the "date" proposal, at time 1 the ORD is NULL, and then at
times 3 and 5 the ORD is t1.  The problem, as we have both noted, is
that t1 is sort of misleading, since there is no way to know whether
t1 really was the original time, or whether there was a previous
example.com with a different ROID.

Under the "counter" proposal, you only get the assertion that the name
is known to have had _at least_ that many previous ROIDs, so at t1 the
counter is NULL and at t3 the counter could be 1 and at t5 the counter
could be 2.  But if the counter fails, it is not wrong: even at t5 it
would not be wrong to be NULL, because the registry might have had a
bug and failed to capture the count.

Best regards,

A

-- 
Andrew Sullivan
ajs at anvilwalrusden.com