Your suggestion to automate the detection of "persistent and excessive"
could work for people and would help "throttle down" those discussions
that need it from time to time, but it would not protect an elist from
spam.
With only one exception of which I am aware (and its not midcom), the
only reason for the "moderation" is to identify spam and to prevent such
one-off messages from ever getting to the subscribers.
Jim
On Fri, 2 Feb 2001, James P. Salsman wrote:
Date: Fri, 02 Feb 2001 01:57:09 -0800 (PST)
From: James P. Salsman <bovik(_at_)best(_dot_)com>
To: l(_dot_)wood(_at_)eim(_dot_)surrey(_dot_)ac(_dot_)uk
Cc: ietf(_at_)ietf(_dot_)org
Subject: Re: Mail sent to midcom (fwd)
Lloyd,
I second your request:
>>... unless you have a specific request for a ... IESG statement,
>
> I'd like a statement that RFC2418 will be adhered to by mailing lists.
So would I. I use multiple email addresses:
[local-subaddr](_at_)bovik(_dot_)org,
bovik(_at_)best(_dot_)com, etc. -- like thousands of other people. And as
people
on this list should know by now, I value pseudonymity and anonymity in
the rare circumstances that they are necessary and in the common
circumstances that they are sufficient.
Lately, the only clear need for this kind of thing has been those
virus-alert email warnings. What's next, computer-prion alerts?
("Warning: This message was edited by the author and not approved
by the U.N. Department of Culture! Further perusal of this message
might eat away at your brain. This message brought to you by a
robot authorized to prevent you from seeing what its creator thinks
you shouldn't." :)
My local USENET newsgroups have a "cancelcritter" that uses a
rule-based system to decide what articles are velveeta. The fact that
it operates behind the scenes is pretty strange. If it would only
summarize the subject lines and source addresses of the messages it
has cancelled on a regular basis, that would be great. But because it
does, some people claim that it often makes mistakes, and so it is
another one of the many similar reasons that cancels are often ignored
by news admins these days.
Similarly, instead of moderating non-subscriber messages, the default
for mailing lists should be to pass them through unless the conditions
described in:
>> http://www.ietf.org/IESG/STATEMENTS/moderated-lists.txt
in particular:
>... 'persistent' and 'excessive'....
are detected. So, for example, if you have the tenth non-subscriber
message in the last hour on a list that usually gets ten messages a
day, then maybe it is time to start holding them for the moderator.
Similarly, for "middle boxes," if you are keeping statistics on the
packets you are forwarding, and all the sudden the proportion of SYNs
from a particular neighbor spikes, maybe it is time to emulate a
source quench on that neighbor. (Or heck, why not even send a few
ICMP source quenches just to say you did.)
And now, for the "thinking outside the box" economic analogy for this
class of problems. Lately, I've been running a data collection
routine that is intended to promote reading literacy using internet
technologies:
http://www.bovik.org/reps-char.cgi
Roughly half of the example children represented in the data presented
by that script are poor readers for their age level. Why are they poor
readers? Because they live in poor school districts with large class
sizes and insufficient insitutional support. Why are they in those
circumstances? Because their wealthy metropolitan neighbors are so
carefully concerned with the education of their own children, that
the often carefully adjust the flow of funds to limit the distribution
based on "performance" such that the schools that already have the
smaller class sizes and the best paid teachers get more money, and no
progress in class size or teacher salaries is made in the
poorly-performing schools.
So, just as some list administrators limit the ability to post in a
timely fashion to those already subscribed, many states have
complicated school funds distribution formulas which act to limit the
resources needed for good education to those who already have them.
In both cases, it is done in the interest of protecting a resource,
ease of communication or reading literacy ability, by hoarding it to
those who already have it.
The analogious solution to the one proposed above would be similar to
the Bush education plan, which only cuts off funds after three years
of poor school performance.
Cheers,
James