> Note  there  is  nothing about spam in there. Nothing about content.
> Nothing  about  consent.  This type of test is different than a test
> that  tries  to  determine  if  a  message  is  spam  and thus has a
> different definition of false positive.

Which is perfectly fine, if this were not a thread about accurate spam
detection  on a list peopled by many overworked and mail-inexperienced
sysadmins of mission-critical MTAs.

As  with  so  many  such  wars, this one began with an offhand comment
closed  with a smiley: RMilner wished we could "figure out how to make
this  happen  to  the  hundred  of  spam  emails  each day and not the
legitimate  emails  from AT&T and Verizon." The solution given to this
query  was  "subscriber  network  filter"--blatantly  counter  to  the
requirements!  (Whether  or  not the problem was supplied in jest, the
solution  appeared  straight-faced  to  me  and  others.)  With such a
mismatch in the context of the between post and reply, why stay within
the  same thread? This list would be more productive if we debated R&D
results  of  experimental  blacklisting measures in their own threads,
rather than mixing-and-matching two very different topics. If not, the
probability   that  sysadmins  will  be  misled  by  the  loop-de-loop
redefinition  of  the  term  "false  positive"  remains too high to go
without comment.

FTR,  though  our  clients  would  not accept single-criteria blocking
without     user-level     *and*     user-controllable    (not    just
admin-controllable)  whitelisting, we are not fundamentally opposed to
it:  our in-house systems actually bounce on a couple of single tests.
But  it  would  be  absurd  for  us  to claim that there are "no false
positives."  It's  an  anti-spam  measure.  It  makes  mistakes  as an
anti-spam  measure.  And  those  mistakes  are,  unequivocally,  false
positives.   Perhaps   there   should   be   an   additional  term  in
there--"designed   FPs,"   "lost-lead   FPs,"   "pimp  hand  FPs,"  or
suchlike--but  they  are  FPs.  On  the  other hand, if you frame your
anti-spam  measure as *not* subordinate to legit message content, then
it's  *not*  primarily  an  anti-spam measure...it can be framed as an
anti-DDoS system, a performance optimizer, a moral guardian, whatever,
all  of which have technical and business merit in their own right and
in  some  environments  overrule  anti-spam  guidelines, but should be
revealed as such.

Overall,  there  continues  to  be an inability for many to understand
what  it's  like  in  jack-of-all-trades  SMB  IT, at a money-crunched
regional  ISP,  and  other  environments. On the flipside, many cannot
grasp  what  it's  like  to  have  the  SLA-guaranteed  freedom,  or a
statutory  mandate, to be as draconian as you want and still keep your
job.  Though  both  are  parts  of the "real" world of IT, I think one
should  err  on  the side of those whose employment is on the line and
explain the perspective of any highly charged recommendation.

-Sandy


------------------------------------
Sanford Whiteman, Chief Technologist
Broadleaf Systems, a division of
Cypress Integrated Systems, Inc.
e-mail: [EMAIL PROTECTED]
------------------------------------


To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html
List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/
Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/

Reply via email to