Re: Turning Razor into a scapegoat for irresponsible mailing list management

Hi,

On Thu, 27 Feb 2003 01:20:39 -0500
Leland Woodbury <razor-users@xxxxxxxxxxxxxxxxxx> wrote:

> Bob,
> 
> Your response focuses on a single point, which is that EFF should be 
> using verification on their mailing list.  This point may be valid, but 
> it seems weak to me, and it misses the point.

Fair enough.

I see two different issues here: first, is Razor vulnerable to
manipulation (addressed below), and why does my particular newsletter
get reported to Razor?

In trying to solve the latter question, I focus on this single point
because it appears to me to be the single most likely reason for Marc's
difficulties and it's something that he can control himself. In order to
find out if someone is maliciously reporting his mailings to Razor, it
makes sense for him to minimize the number of ways someone can
unintentionally report his mailings to Razor, provided the changes to
his system are reasonable and within his sole control.

[snip]
> Someone made a similar point earlier in this thread:  Razor's inherent
> limitation (as well as its strength) is that it's a group effort,
> whereas spam is in the eye of the (individual) beholder.
[snip]

I agree, but I took Marc more literally. I'm less worried about people
who flag legitimate mail as spam because they occasionally forget
they've signed up for something or use definitions of spam I don't agree
with. I expect these casual anomalies to be background noise that will
be drowned out by the voice of the masses, either from overwhelming
reports of a message's spamminess or by the drop in trust following
report revocations by others. That's not censorship, that's consensus or
at least majority rule. In this case, nobody is specifically trying to
silence EFF.

My concern was the case where an organized group could use Razor to
intentionally interfere with the delivery of legitimate mail as part of
a political, social, or religious agenda[1]. I believe it can be done,
that it's difficult to do and maybe more difficult to defend against.
Conspiracies of this magnitude are difficult to keep secret.

> And Marc needs to understand this:  The Razor system is working as 
> designed, and that's what's causing his problem.  And incidentally, I 
> don't think that validating emails will necessarily make that problem go 
> away.

I agree, although I believe that confirming his list will reduce and
maybe even solve his problem. We can comfortably disagree on this point,
because until his list is confirmed and we have data before and after,
it doesn't really matter.

IMHO, it hurts EFF little to observe mailing list management 'best
practices' like confirmation, bounce handling, etc. Regardless of what
confirmation does to their Razor rating, it makes their system much less
abusable. Virtually every mailing list I'm on requires confirmation[2],
every list I maintain requires confirmation, and I don't see why
expecting EFF to confirm their mailings is so egregiously offensive,
especially when operators of much larger lists seem to have little
trouble doing it.

Marc has steadfastly dodged[3] any discussion of running EFFector as a
confirmed list, continues to assert that there's something wrong with
Razor and that the burden of proof should be on Razor to prove it isn't
being used maliciously. I find it irritating, immature, and damaging to
EFF's credibility.

> Back to my ignorance about the TeS:  Can you or anyone else explain 
> precisely how this works?  The TeS assigns a weight 0-100 for each of 
> us.  How are weights assigned exactly?  Is it enough for a single 
> reporter with a strong weight (or a small number of such reporters) to 
> force a message to be spam if no one revokes his report?

The TeS algorithm is a mystery to us laymen. This causes me minor
indigestion but I can live with it since I use Razor in combination with
SpamAssassin, buffering out any transient weirdness with Razor. Others
who drop or reject mail solely on Razor results should lose some sleep
over this (again, see [1].)

Thank you for civil disagreement, a cogent argument, and for not calling
me a doody-head.

-- Bob

[1] In the same way that most web filtering programs keeps their
filtering scheme hidden ("trade secret") and block a lot of sites that
don't match their overt agenda. One could draw parallels between the
'black box' nature of the TeS system and the encrypted list of blocked
sites for NetNanny, et. al., but if so, wouldn't we hear of a lot of
other lists that were being blocked by Razor? Is EFFector the only list
that has this trouble with Razor, is it only unconfirmed lists, or are
there lots of lists affected and we're only hearing from EFF? I'd expect
NARAL, NOW, ACLU, and a whole host of advocacy groups to be affected if
there was a conspiracy afoot. Since we don't, and since I can show a
simple way EFFector can be reported to Razor with no malice or
conspiracy on anyone's part (as well as a simple fix for that issue), I
seriously doubt there's a conspiracy to silence EFF using Razor, at
least not one with a reasonable chance of success.

[2] Over 20. Some are very small with very low traffic where I know the
operator personally. Here, confirmation is not so important.
Incidentally, this is meant as anecdotal evidence, not proof.

[3] 'steadfastly dodged'? How do you do that? Reminds me of the part of
'Raising Arizona' where the bank robbers tell everyone to simultaneously
freeze _and_ get on the floor. Sure, the phrase sounds good but
literally, it's stupid. :/


-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com