From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Fri Sep 28 2001 - 05:25:14 CDT
In <20010927130802.F17334@main.templetons.com> Brad Templeton <brad@templetons.com> writes:
>On Thu, Sep 27, 2001 at 03:53:39PM -0400, Forrest J. Cavalier III wrote:
>> > As for Clive's suggestion that we should now discuss cancels a little
>> > more, I entirely agree. The simplest way to implement multiple cancels
>> > would be to allow multiple message-ids in a single cancel message
>> > (indeed Henry already implemented that in CNews in a fit of enthusiasm)
>> > and it would hardly require rocket science to incorporate it into other
>> > serving agents.
>>
>> Is this intended to be a reinvention of nocem?
I think that is an issue we should seriously look at. Not for
implementation in this draft, but it would be useful to know what we had
in mind for the Security Extension that is to follow, so we can slant our
present text if necessary.
>However, in practice, it might often be the case that the attribute
>would be a recommendation for cancel, which would indeed be acted upon
>immediately if the sender is trusted, and as such not really be stored
>in a database.
Exactly. The NOCEM has a well understood syntax, though it could maybe do
with a bigger set of attributes, including perhaps "recommended for
cancel". Indeed, if properly done, one might be able to do away with the
present system of bulk cancels (with its huge waste of bandwidth) and rely
entirely on NOCEMS for spam fighting. If we went that route, we could
include a proper definition of the NOCEM in our Security Extension, and
the Cancel control message could revert to its originally intended use.
It seems to me that the only other alternative is to strengthen the
existing Cancel control message, but that raises interoperability issues
with INN and other existing software.
I think we should now discuss which of these routes we wish to follow, or
whether there are any other options. Perhaps some discussion on
news.admin.net-abuse.policy might be useful too.
>However, at this point, I don't want to complicate matters so would not
>push hard for the generalized design, though I tend to like them. However,
>the format should be sufficiently generalized.
I think the biggest problem with NOCEMs at the moment (and Cancels share
this problem too) is the distribution. Nobody can afford to subscribe to
news.lists.filters unless he is running a large site with a full feed. The
alternative may be to crosspost them to the groups affected, but then they
will appear within the traffic on the group, which some people may not
want. However, that sounds like a problem which could be solved, either by
the use of a Distribution: nocem header, or by killfiles, or something.
-- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl Email: chl@clw.cs.man.ac.uk Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K. PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5