[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: List spam cleanup: half-time scores



Quoting Frans Pop (elendil@planet.nl):

> Holger and Christian have already started the reviews for the next two 
> years: 2004 and 2005.
> 
> Detailed info is available from:
> http://wiki.debian.org/DebianInstaller/SpamClean


Small update:

2005 has been processed by Holger Wansing and myself (plus 3 months
done by Don Wright additionnally). So, we need 3-4 people to work on
these months so that more spam is identified for reviewing.

I completed 2004 (the biggest year ever, for D-I, sometimes with up to
4000 messages a month), and Holger is on his way completing it as
well. It will take time to process these months. I actually wonder
whether we could ask listmasters to lower down the bar for years
before 2005 or so. Otherwise, knowing that this is huge work is likely
to discourage everybody.

About the spam storm of August 2008, I proposed Cord to remove those
mails but got no answer (or missed it).

I just completed 2003 which is *much more* easier as months have about
500 messages for the first half of the year and about 1000 for the
second half. There is "few" spam in 2003 (but stiil between 20 and 40
each month).

I'll continue going backwards.

My current conclusion is that grabbing the list archives as mailboxes,
processing them through crm114 and then browse then in mutt is *very*
efficient. I actually completed 2003 in about 3.5 hours of quiet train
travel.


About signalled spam reviews: as of June 21st, I was up-to-date for
-boot but very few new spams were marked between June 14th and
21st. This is certainly the consequence of processing work being done
by Holger and myself only these weeks. I still have to process the
result for June 28th but I don't expect much more signalled spams.


Attachment: signature.asc
Description: Digital signature


Reply to: