Dave Crocker <dhc@xxxxxxxxxxxx> wrote: > > The difference is that there are practicalities of implementation and > use that we have to anticipate. This falls under the unfortunate > reality that the real-world is not conducted so carefully. I have great respect for Dave's viewpoint on that issue. But I do think there's a principle here that doesn't depend upon the implementation: that silently dropping a false-positive _does_ create problems as perceived by the end users -- and that those problems would be significantly reduced if the innocent sender of the false-positive email were notified of the failure to deliver. > On the average, user-level Internet mechanisms need to be pretty > simple and straightforward, if they are to be successful. Omigosh yes! I've taken far too many support calls: "Is the Internet down?" to think otherwise... But I, at least, am thinking in terms of an implementation where we notify the SMTP-sending-server during the SMTP session, with a message including a URL for more information. IMHO, this would tend to converge to a situation where end-users understood the issue -- and learned to route around it. ;^) -- John Leslie <john@xxxxxxx>