On Sat, Oct 23, 2004 at 07:18:57PM +0800, test wrote: > >The advantages of the new technique: 1.As a receiver,you first >judges the useful of the email by simple >information(email-pointer:subject,from,to and etc).if is,you can >refuse it to forbid download the body of the email(reducing the >Internet flux) This is not an advantage over existing SMTP. Case 1: If you can judge by the given simple information (basically the mail header), then you could simply discard/tag the message after reception or even abort the connection after transmission of the header. (Violation of SMTP, but spam is an excuse). So there is not advantage over SMTP. Also no advantage from user's point of view. If you can tell from sender/subject that it is spam, than you won't read it. Just a bandwidth matter. Case 2: If you can't judge, then you need to fetch the message anyway. Again, no advantage. This is also a legal problem: What is the transmission time? Imagine you have to fulfill a contract by sending something by email. When did you fulfill? when the receiver starts to fetch? How long would you have to keep your server online? A day? A week? A month? The proposal is not new, and has been discussed on ASRG about more than a year ago. You furthermore run into several problems: When will the server be allowed to delete the message? After a download? What if the message is aliased to several recipients? How would the sender tell how many downloads it will take for all recipients to get the message? What if the recipient never downloads? What if the diskspace of the server is exhausted? It is also a security problem: You need to protect the server against faked fetches. E.g. if a message is forwarded or bounce to someone else, the new recipient could delete the message from the server. BTW, it is not correct to assume that all people use pop3 to fetch messages. How should someone fetch emails when not online? (e.g. I do use UUCP). And, ironically, it makes spammer's life even easier: Your proposal makes sending email much more complicated for those who deliver normal mail. But for those who do mass mailing and are sending the same message a million times, your proposal saves a huge amount of bandwidth, because they need to transmit the header only and to keep just a single message available for download for those who read the message. So it just reduces the bandwidth needed for spamming. That's not exactly what an anti-spam-system can be expected to do. regards Hadmut _______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf