[Asrg] Re: Receiver Initiated Authentication

[SM]

At 20:32 17-09-2007, Michael Kaplan wrote:
>Section 9.3 calculates how RIA, when adopted by 50% of the global 
>population, will on average increase 'spam' sent to innocent users 
>by 0.2%.  Unlike with real spam BATV can perfectly block this 0.2% 
>increase in spam received.  I doubt that RIA will be the straw that 
>breaks the back of the bounce system.

Your proposal mentions bouncing "difficult to classify email".  It 
may not be the straw that breaks the system but it may prompt the 
adoption of measures to block that type of email.

>Who will pay for it?  I expect that the client and server software 
>developers will take it upon themselves to incorporate Auto-Resend 
>as a routine feature.  It is a one-time software upgrade.  It is not 
>like a statistical filter that requires constant maintenance.

The problem with every new proposal that is tabled is the cost 
involved and deployment factors.  One may see it as a one-time 
upgrade but it's non-trivial to implement because it involves either 
updating the MTA or the MUA.  Software developers are usually market 
driven especially when it comes to MUAs.  If they implement your 
proposal, it will only be available in new versions of the software.

I suggest that you implement the system you propose and publish the results.

Regards,
-sm