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

On Oct 08, Gert Doering <gert@space.net> wrote:

(But I've said this before :-) - I do not see it as a useful excercise having to create an organization: object for the sole purpose of being able to have a different abuse-c: for some inet(6)num object) Agreed. I am not sure about who actually disagrees with this among operators.

-- ciao, Marco