> From: Keith Moore <moore@cs.utk.edu> > ... > > That is the same sort of reasoning that spammers use to justify > > their activities. > > it's also the same sort of reasoning that you used to justify > sending this mail to the IETF list. You know that is a false statement. You know that subscribers to this list ask to receive all of the good stuff with the drivel, except for the out-right spam and the messages from banned individuals. You also know that no one asks to have their SMTP servers do a useless cycle in addition to receiving an auto-response. I do agree I send too much mail to this list. On the third hand, you have me and everyone not in my filters beat by a factor more than 10. > > - bounces would not bog down your mail servers, since you can discard > > or otherwise deal with bounces with no more "bogging" than you are > > spending poking at other people's systems. > > for a variety of reasons, that's simply not true for the cases where > I'm using it. How did I guess you would see things that way? It is, of course, nonsense, because you could arrange to not receive any bounces just as some spammers do. > > (Many spammers want to receive bounces to clean their > > lists, because some big ISPs and others automatically black-list > > SMTP clients after they've sent to too many bad addresses.) > > perhaps, but even the spammers who clean their recipient lists don't > necessarily use valid return addresses, and it's the return addresses > that I'm checking. That's nonsense that you would see as such if you were not stuck on rationalizing that which you know is dubious. Spamemrs who want bounces to clean their lists do indeed use valid return addresses. > > - a mechanism that could really determine that addresses are valid > > could be useful in a web page to provide immediate feedback to > > users. The reality of MX servers, firewalls, sendmail "catch-all" > > maps, and other things make the RCPT command too unreliable to use > > for that purpose. > > BS. a simple syntax check isn't perfect either, but it's useful to > catch some errors. this catches some more errors, without exhibiting > false positives. it also has a cost, which is why I haven't > recommended it for general use. but that doesn't mean it is useless. Charges of "BS" from someone who admits never having measure the effectiveness of the mechanism are ironic. I'll shut up now about this. Vernon Schryver vjs@rhyolite.com