Hi, 

For reasons I am still investigating, I did not receive razor-users mail
for the last two weeks. I just caught up on the censorship thread, which
I believe is rapidly losing its signal/noise ratio. Shane Williams
recently posted a nice clarification, and I'd like to injerject a couple
of points too:

1) razor-check implements the concept of "minimum confidence" which is a
   threshold that needs to be met before something is considered spam. The
   razor servers recommend a small min_cf value which is computed as the
   average confidence the average user of Razor2/SpamNet is comfortable
   with. However, users are _NOT_ restricted to use this value. If the
   mailing lists you subscribe to get regularly filtered by Razor2, please
   change your min_cf to a higher value -- like 50. In the Razor2 trust
   metric, confidence corresponds to consensus. cf of 100 means 100%
   consensus in the community, a cf of 0 means 0% consensus. If you have a
   mail stream that is heavy on newsletter traffic, I'd highly recommend
   increasing your min_cf value in .razor/razor-agent.conf

2) The upcoming SpamNet upgrade has a new feature called "auto-unblock"
   that learns from people's behavior and automatically unblocks
   newsletters/lists people have chosen to receive, soon as they receive
   them. This is going to have a huge impact on newsletter false positives
   for the Razor2/SpamNet community. We also plan to enable preemptive
   revocation when we release the new clients. Unix users can achieve the
   same functionality by auto-revoking newsletters using procmail or mail
   processing system of their choice. Either way, Unix users will benefit
   from the auto-unblocks from SpamNet clients.

3) I have immense respect for EFF, what they stand for and their
   achievements. Whenever free speech is threatened on the Net, EFF is
   there to protect it. We can't be thankful enough to them. I would be
   _really_ hard pressed to believe they send unsolicited mail. If their
   legit mailings gets blocked, their frustration is only to be expected.
   I am surprised by the anti-EFF sentiment in this thread which seems to
   be backed by random guesswork. I not interested in pushing my personal
   opinions on people, but this list is definitely not the place to voice
   your concerns about EFF's mailing policy. You should talk to the good
   folks at EFF if you have any question or concerns.

4) I am not familiar with what MoveOn is or does, whether or not they have
   clean lists, whether or not they deserve to be blocked by Razor2. Their
   political philosophy and details of operations, IMHO, are outside the
   charter of this list. If they do have a lot of legit subscribers, the
   confidence value of their mailings will be low and the real subscribers
   can receive their mailings if they bump up the min_cf. I believe
   auto-unblock will have a good effect on their list if it is indeed
   purely solicited.

5) There are many reasons why users report newsletter mailings as spam.
   Very often the users are too lazy to unsubscribe, and sometimes the
   unsubscribe mechanism is too complex. Such users might do some harm in
   the short run, but eventually the trust system will rate them down.
   Razor2 is an emergent system that improves with every report and
   revoke. There are several newsletters that used to get marked in past,
   that don't get marked today because those who were reporting them have
   been sufficiently penalized.
   
6) Claims of censorship are, well, greatly exaggerated. Censorship is
   about coercion, and there is no coercion involved here. Razor2 is
   simply a system that publishes the opinions of its users. The
   technical challenge for Razor2 is to communicate the collective
   opinions effectively, and this is something we keep improving all the
   time. In near future, there will be changes to the trust metric that
   would communicate more granular values of consensus and disagreement.
   Razor2 does not and will not inflict its own opinions on you. It
   doesn't have any.

7) Accurate spam filteration is a hard problem, and even though we do a
   good job of it today, there's a process of evolution associated with a
   system like Razor2/SpamNet. This process can be frustrating to some
   users whose mail habits are different from most as well as to legit
   mailers like EFF. I would request everyone to recognize this and cool
   down instead of taking pot shots at each other. Aggresion, in general,
   is counter-productive and not welcome on this list.

cheers,
vipul.

-- 

Vipul Ved Prakash          |   "The future is here, it's just not
Software Design Artist     |    widely distributed yet."
http://vipul.net/          |              -- William Gibson


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Razor-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/razor-users

Reply via email to