@Kilgaard

> Anything that makes it harder for SPAMMERS to SEND email looks good to me.

Unfortunately DKIM makes it much harder for legitimate senders than for spammers. If hitting spammers means killing email, what is the point? If you follow your point, then we should move from email to proprietary, secured protocols. Exactly the dream Bill had for many years. The challenge against spam is to make it hard for spammers but let legitimate email thru.

>Of course, somebody will probably point out that the CPU costs for bot-spammers > is almost zero anyway because they are just using their zombie hosts CPUs.

Exactly. Cost for spammers is 0. Sending emails via gmail accounts created by hijacked zombie PCs costs 0. And the emails are DKIM / DomainKeys signed from gmail.com! You want to block all emails with a valid DKIM signature for gmail.com domain? It will certainly make it harder for spammers.

Bill because if he had it his way, smtp would not be used anymore.

@@ spf (by Olivier)

> The big issue ( imho ) with them are:

>

> a) implementation costs for sender.

>

> b) cpu costs for the sender. If you send many emails, it is very expensive in terms of ressource to compute these signatures

>

>

Anything that makes it harder for SPAMMERS to SEND email looks good to me. The major cause of spam is that it is just too easy/cheap to send spam. If the protocol causes additional cost (even in terms of cpu load) to the sender of emails then this will greatly impact the profitability of spam sending.

That this cost would have to be carried by legitimate email senders also is unfortunate, but a necessary price to pay.

Of course, somebody will probably point out that the CPU costs for bot-spammers is almost zero anyway because they are just using their zombie hosts CPUs.