On Monday 11 September 2006 21:29, J.D. Falk wrote: > On 2006-09-11 17:05, Hector Santos wrote: > > ----- Original Message ----- > > From: "J.D. Falk" <[EMAIL PROTECTED]> > > > >> Yahoo! will use Yahoo!'s internal systems to make our own internal > >> decisions about each message. AOL will, I'm sure, use AOL's. > >> Everyone will make their own decision in their own way, just > >> like today -- perhaps with a 3rd party's input, perhaps not. It's > >> up to them. > > > > No doubt. Just like today. > > > >> A DKIM verification result is just one more input. > > > > In what way? > > IP address is an input. Each URL in the message is an input. Virus > scanning results on attachments are an input. Filtering hasn't been > binary for years.
While that's true, I think part of the point is to limit the need to filter. For example, mail to my domains is scanned pre-queue and rejected if it contains a virus or is from a client on certain reasonable safe RBL lists. That's binary and it's not filtering, so it isn't at all counter to your point. What's left goes into a filtering process that is, as you say, not binary. I'd like to get rid of a larger fraction of the 'bad' mail in the rejection stage and deal with less of it in the filtering stage. Accept/reject criteria that can be used add to the reliability and consistency of the e-mail system. I do think that the SSP idea has potential to add value within the filtering regime too. I would imagine that a message from a 2822.From domain with no signature and no SSP and a message from a 2822.From domain with a 'sign-all' SSP and no signature would be weighed differently during the scoring process. Even if you are just in the filtering business and not in the rejection business, SSP gives you another piece of information to add to the matrix. Scott K _______________________________________________ NOTE WELL: This list operates according to http://mipassoc.org/dkim/ietf-list-rules.html
