[Groff] RE: List delays

[Ted Harding]

On 22-Feb-06 Ted Harding wrote:
> It seems, looking at the headers of recent mails to
> the list, that postings to the groff list are being
> delayed at lists.gnu.org by 6-12 hours, typically
> over 9 hours.
> 
> Is this (as I hope) a temporary phenomenon?
> 
> Best wishes to all,
> Ted.

With Miklos's and Werner's messages of

  Feb 22 07:24 GMT distributed to me 18:24 GMT
  Feb 22 07:28 GMT distributed to me 18:41 GMT

the delays at lists.gnu.org are now at least 11 hours.

Looking at the list archives at

http://lists.gnu.org/archive/html/bug-groff/2006-02/index.html

I see that the archived messages for the last 7 days are 95.5%
spam (some of which is of the well-known groff speciality).
Specifically:

            Spam  Real   All
Wed 22 Feb     2     0     2
Tue 21 Feb    25     0    25
Mon 20 Feb    24     0    24
Sun 19 Feb    16     4    20
Sat 18 Feb    12     2    14
Fri 17 Feb    30     0    30
Thu 16 Feb    18     0    18
----------------------------
             127     6   133

Does this have anything to so with the delays? Is lists.gnu.org
choking on this stuff?

I am concerned about this, but admit that a solution is not
entirely trivial.

I run a couple of mailing-lists (man-lug and linux-users
@lists.manchester.ac.uk), which are of course targeted
by spammers but at a somewhat lower level (on average
less than 5 spams/day each, but up to 12 or so at times).

The lists are configured to accept postings from subscribers
only. All other postings are notified to me for approval or
other action ("discard" in the case of spam).

This involves a certain amount of work (though not much):
notifications which refer to spam are easily identified and
I leave these alone till I feel like dealing with them.

Maybe once or twice a week a genuine message comes through
from a non-subscribed address (sometimes from an outsider
with an interesting and relevant communication, sometimes
from a subscribed user posting from a non-subscribed address).

When that happens, I then visit the list admin webpage, approve
the genuine message[s], and take the opportunity to flush away
the spam.

Also, both lists have their archives accessible to subscribers
only (password required). This may help to reduce the spam somewhat,
but I am not sure about that (though it certainly helps protect
list-members' addresses from spambots).

I'm not sure about the desirability of making the groff list
subscribers-only, given that one would like to promote open
acess! But things can reach a level where the issue has to be
faced.

Probably restricting reader access to the archives is very
undesirable. The anti-spam effect is likely to be minimal,
while the impact on open access would be severe.

I note, however, that the groff messages which I actually
receive are (except for those few notorious ones "sent" by
a few of us, myself included) are non-spam. At least, that
is the case where there is "[Groff]" in the subject. But
the spam in the archives is otherwise 100% without "[Groff]".
So maybe there is already a mechanism in place for Werner
(or someone) to check out suspect messages; but in that case
why do they get into the archives?

Or maybe some of the multitude of spam which I receive (but
never look at) is also sent out by lists.gnu.org but without
["Groff]" in the subject. I'm a bit perplexed by my observations.

Just raising some thoughts for consideration! (And hoping
that the gridlock eases enough for discussion to begin to
flow again).

Best wishes to all,
Ted.

--------------------------------------------------------------------
E-Mail: (Ted Harding) <address@hidden>
Fax-to-email: +44 (0)870 094 0861
Date: 22-Feb-06                                       Time: 20:06:28
------------------------------ XFMail ------------------------------