logo       

Misconceptions: msg#00203

mail.spam.razor.user

Subject: Misconceptions

In several recent threads, I've noticed what seem to be some
misconceptions about Razor (or I've got some misconceptions about
razor). I'll try to list and correct a few here, but if I miss
important ones (or get them wrong!) please add. My point here is not
to make anyone feel stupid. Rather, if we're going to have a
reasonable discussion, we need to figure out what basic facts we agree
on and which ones we really differ over.

1. Razor is doing x (where x has recently been posited as censoring,
but could be and has been other things)

In my mind, Razor doesn't do anything. Razor is the name given to
a logical system whereby an arbitrary number of people can submit
to and query from a database. This database holds hashes
identifying fuzzy versions of specific emails that are submitted
by someone as meeting that someone's definition of spam. In
addition, there is a trust system that maintains data about
submitters and how their submissions compare to other submitters
submissions. (When I use the term submit, I mean either report or
revoke) I hate to compare the two, but guns don't kill people,
people kill people.

2. Razor says x is spam (where x is the body of some email)

Note first that x above is not some source of email such as the
domains EFF or whitehouse.gov or makemoneyfast.com or IP address
or an email address. x is only the body of some email. While
someone with a spamtrap may effectively report all email from a
given source as meeting his definition of spam, there is an
important difference.
More to the point, Razor itself has no definition of spam. Only
submitters and queriers have definitions of spam. Submitters'
definitions are reflected in what the report and revoke.
Queriers' definitions are reflected in their minimum confidence
level (or in the case of SA users, the point value they apply to
razor).

3. If x is listed in Razor, then at least one person thinks it is
spam (again, where x is an email body)

While Razor is designed such that a report (as opposed to a
revocation) represents someone thinking it is spam, there is no
necessary limitation on the actual thinking of the submitter (or
for that matter the existence of a submitter capable of having
thoughts at all). x may be listed in Razor for a whole variety of
reasons having nothing to do with the perceived "spamminess" of x.
All the above is also true for the process of revocation.
The point here being that while Razor was designed as a tool to
create a collaborative database of spam, there's nothing stopping
others from misusing it or using the tool to do something else
entirely

4. x is in the Razor database because of reason y (where x is the
body of an email and y is any reason other than someone or
something submitted it).

As someone else pointed out, and point 3 gets at (but bears
repeating), there is no way of knowing the motive (or reason) a
given submitter sent x to the Razor database (or again, that the
submitter is even capable of motivation). As there is no way of
obtaining an anonymous person's motive, the only way to know is if
the person tells you, and even then, since it is anonymous you have
no way to know that the person offering their motive is the person
who even submitted in the first place. Anyone else positing a
reason other than "because it was submitted" is just guessing.
This includes you.


That's all that comes to mind this instant (though I know there are
others).

--
Public key #7BBC68D9 at | Shane Williams
http://pgp.mit.edu/ | System Admin - UT iSchool
=----------------------------------+-------------------------------
All syllogisms contain three lines | shanew@xxxxxxxxxx
Therefore this is not a syllogism | www.ischool.utexas.edu/~shanew



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf


<Prev in Thread] Current Thread [Next in Thread>
Google Custom Search

News | FAQ | advertise