[db-wg] abuse-c + org / inetnum

On 7/10/13 17:04 , Niall O'Reilly wrote:
> 
> On 7 Oct 2013, at 15:02, Gilles Massen wrote:

>> Apparently the only workable solution at this point is to create a
>> duplicate organisation object with a different abuse-c.
>>
>> Is data duplication really the solution that the WGs would like?
> 
> 	I'm glad you've re-opened this question, Gilles, as I just
> 	last week had occasion to update an object which had long
> 	had multiple 'abuse-c' contacts, and had to choose which one
> 	to keep.
> 
> 	I'm in the "No" camp wrt your question above.

Glad to hear that. I'm not sure I would make use of multiple abuse-c for
a given object, unless one could provide some additional information
with them.

> 	Back in the original "abuse-c" proposal [*],
> 
> http://www.ripe.net/ripe/mail/archives/db-wg/2004-January/002489.html
> 
> 	there was provision for multiple "abuse-c" attributes,
> 	but this was not carried over into the current specification.
> 
> 	Would "hint strings" be a way to go, or have you something
> 	else in mind, or are you just re-opening the question?

The short version: No, yes, not really :)

A bit more verbose: While I certainly would welcome a mechanism like
'hint strings' and have use for it, it would not apply my case where I
simply need different abuse-c for different subnets, not due to the type
of report but rather for the people that should read/react upon them.
Not all inet*nums are equally important.

As for re-opening the question: I feel that it was never closed, only
abandoned over summer time. But the most important driver for coming
back is that the only available solution (data duplication) is so
tremendously wrong that I'd really like the WGs to be conscious about
it. It the members feels that's ok and I'm only paranoid about data
quality - fine. Silence, however, it not going to convince me.

Something else in mind: as before: allow abuse-c for inet*num. Prefer
and encourage the organisation way, but allow the other. Even if I did
share Tobias' belief that attaching abuse-c to inet*num would weaken
it's overall quality, I'd always prefer a hypothetical issue
(intenum+abuse-c) over a certain one (multiple identical roles), unless
the former is substantiated by something more than 'someone might get it
wrong'.

If anyone has a different idea, please share!

Best,
Gilles