Isnt it already in place to lower outlook scores? Why not just increase the
score for HTML mails and increase the amount deducted for outlook?

-----Mensaje original-----
De: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]]En nombre de Bob
Proulx
Enviado el: Monday, January 20, 2003 5:12 PM
Para: [EMAIL PROTECTED]
Asunto: Re: [SAtalk] HTML only messages


Jeffrey Culverhouse wrote:
>
> I know the scores are tuned from a large directory of email, but I was
under
> the impression that HTML ONLY emails are very often spam and wondered what
> the consensus of this list was??  Right now, it looks like it is just 0.4,
> but I was thinking of raising that to 2.0 or so...  I know, I can do what
I
> want, I just wanted a list opinion...

Traditionally in the technical community the present of HTML was a
strong indicator of spam.  The original spamassassin rules reflected
that and the scores for any html were quite high.  I am in the
technical community and html is not a suitable email format.

But then SA because very popular and use spread to the business
community.  The business world, it seems, thinks that html email is
great and should not be counted as spam.  There were many sample
emails submitted to the genetic algorithm engine which creates the SA
scores.  Therefore later versions of SA rules score html mail
neutrally since in the business world it is not an indicator of spam.

The primary difference and reason for divergence between these groups
is that technical folks see the problems with MS-Outlook and generally
refuse to use it.  Business folks use MS-Outlook regardless of the its
bugs.  MS-Outlook is arguably one of the biggest sources of problems
which water down spam filtering.

> How many people running SA override the scores of particular rules and is
> there some way of identifying rules that need higher scores (and most of
you
> are overriding) and really need to be adjusted in the distribution, if
> any???

I override the present low scoring rules for html and increase them to
a larger value.  But I am rather contained and still the points added
for html is lower than I would prefer.  Here are two of my rules you
might find useful.  These are in addition to any SA added score for
HTML.  YMMV.

  header HTML_MESSAGE           Content-Type =~ /text\/html/i
  describe HTML_MESSAGE         HTML message
  score HTML_MESSAGE            1.2

  body HTML_IN_BODY             /Content-Type:\s*text\/html/i
  describe HTML_IN_BODY         HTML in body of mail
  score HTML_IN_BODY            1.2

The new Bayes inference engine being added to the development version
of SA has potential to learn this difference between communities.  It
will automatically determine that any html mail to me is spam but that
to my business friend html mail is no indicator of spam one way or the
other.  I feel certain that it will do the job well within a given
community.

But being successful there may cause a bigger chasm between these
groups since they will still be speaking differently.  That is, I will
"email them" in plain text mail and they will "outlook me" in html
encoded mail.  They will be able to read what I say but they will not
be able to communicate to me.  Hmm...  Perhaps that is a good thing
after all.

Bob




-------------------------------------------------------
This SF.NET email is sponsored by: FREE  SSL Guide from Thawte
are you planning your Web Server Security? Click here to get a FREE
Thawte SSL guide and find the answers to all your  SSL security issues.
http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0026en
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to