[acm-tf] Draft of RIPE NCC Fuctionality Document

Hi Brian and all,

On 13/Jun/11 17:06, Brian Nisbet wrote:
> In conversation with Denis Walker from the NCC he let me know that he
> had a document which proposed a way of dealing with some of the
> matters the TF is dicussing. I'd like to share this with you today
> (and, indeed, apologies for not sending it out to you sooner) and
> discuss it on Wednesday. The document is a design approach based on
> the current DB model that could form the basis for a policy
> discussion.

I think working out this kind of stuff is exactly what a mailing list
is optimal for doing, as participants can propose snippets of text and
give +1/-1 feedback about them in order to reach consensus.  If Denis
had posted this file to the list in May, we could have begun
discussing it then.

> I am by no means suggesting this is the only approach, but I do
> like what is laid out in the document and so I submit it to the TF
> for discussion.

I like it too, and have some comments.

Foremost, the paper has a section named "What to do if no dedicated
abuse handler found", but never actually spells what to do if it /is/
found.  It obviously assumes people have (their own) knowledge about that.

The document mentions "a web interface for the public to use
directly."  I hope it doesn't refer to end-users.  IMHO end-users
should just click This-is-Spam buttons.

I think the tool will mostly be used in scripts (for forwarding
messages reported as spam to the right handler(s) --a process to be
defined).  An interactive version would be fine, of course.

Care has to be taken for the "e-mail" attribute not to be misunderstood.

Abuse reporting is going to be a cooperative effort.  Defining abuse
contact to be mandatory may result in abuse-mailboxes directly
connected to /dev/null, a waste of resources for both setting up and
running them.  Having TiS buttons directly do nothing in such cases
would sort the same effect more cheaply.  OTOH, spam scoring tools may
learn to take into account whether a message can be reported as spam
or not, so I'd rather opt for auto-removal of non-functioning entries.