From: "Geo." 

Yes I am parsing the html email messages to distinguish visible text from
raw text. That's only one aspect of the filters, I also have the capability
to parse the raw email.

See http://www.gordano.com/home/press/releases.htm?NewsID=2087&Next=One

So given the choice to filter a specific spam by a phrase that almost never
appears on normal emails or to find some feature that lots of spams have in
common but normal email almost never have, which would you use?

Geo.

"Randall Parker"
 wrote in
message news:3f8f59d7$1{at}w3.nls.net...
> Geo,
>
> Your problem is that you are not parsing the HTML email messages to
> distinguish visible text versus stuff that is not seen by the users. If
> you had a library call that you could pass an email to that would return
> just the visible text you could do your checks on that.
>
> There is probably such a call available from cpan.org for Perl. Can you
> write filters in Perl? Or get a Java lib or a C++ lib that will do it.
>
> Geo. wrote:
> > This is the kind of bs that really shakes my pop..
> >
> > You know how spammers are using html comments like
 to
> > break up phrases and words like "viagra" so they can
get their lame
spams
> > past the filters? Well not to be outdone I came up with a filter that
counts
> > these comment lines and after 8 it decides this is spam. Works great,
right,
> > I mean who in their right mind hand writes html email code and comments
the
> > code?
> >
> > Well Microsoft to the rescue.. All the idiots who use Word 10 as their
email
> > editor are now generating code with tons of comments like:
> >
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> >
> > I mean come on people, is anyone at Microsoft even thinking about how
this
> > affects things, do they WANT more spam? What are you guys thinking,
someone
> > actually reads the SOURCE besides email admins trying to filter the
crap?
> >
> > Geo.
> >
> >
>

--- BBBS/NT v4.01 Flag-5