On Mon, May 23, 2005 at 12:07:23PM +0200, Shane Kerr wrote:
Andre Koopal wrote:
Last week we updated all our allocations with an IRT object.
However, this ended out to be quite useless as the IRT object isn't shown in a default query, which 99% of joe-user will do, they don't know to specify -c.
Another problem I noticed is that the irt-nfy isn't suppressed if you do specify -c.
So I want to propose to show the IRT object in a default query and to also hide the irt-nfy attribute in that default query.
I definately think that the "irt-nfy:" attribute should be hidden - this was basically an oversight with the abuse change proposal on my side. The "upd-to:" and "mnt-nfy:" should also be filtered.
As for the "-c" flag, I would propose that the default behaviour should be:
1. Perform a lookup that works the same as "-c" today, and if it would return some INETNUM objects, then finish (returning IRT objects as today)
2. If no INETNUM objects would be found, fall back to the current behaviour
We could have the "-c" flag change its meaning, to give you the query without consideration for the "mnt-irt:" attribute, just like today.
APNIC has a model where members can hide information about their assignments:
http://www.apnic.net/news/2004/0930.html
By implementing the "-c" change above, the maintainer managing the Whois database records int the RIPE Database could decide which block users would get back on a query. I think that this would give the benefits of the APNIC approach (for me, getting users to the "best" person to help them with their problems), without what I consider a drawback of the APNIC approach (there is "secret" data in the database).
A futher suggestion that Wilfried (I think) mentioned to me, was to add "mnt-irt:" to the AUT-NUM object type.
As I don't see any more comments on this, have we already reached consensus? Wilfried: how to best proceed with this? If there is hardly any discussion about this, it seems silly to wait for a RIPE meeting on this. Regards, Andre Koopal MCI