Dear Mr. Haskins

Is your patent application related to the technologies you discussed in the article?

Regards, Ivan.

While any property of an e-mail message (including a checksum or the presence of certain keywords) may be useful to the recipient for determining a reputation score, many of those properties can be tweaked by the sender with few or no limitations. Only a few items such as the SMTP client IP adress can be relied upon as authentic.

Besides rejecting unwanted mail, a carefully crafted reputation system may be used also to send a strong message to any network provider hosting spammers that their presence isn't appreciated. A checksum has neither the incentive nor the means to improve its own reputation, but a corporate entity should seek to protect its reputation as its own trademark, and prospective customers should review it before signing up in order to learn whether their e-mail is likely to be delivered or rejected.

For this reason, I'm inclined to rely exclusively on address-related information for spam prevention. Even if I could easily set up my server to reject messages containing certain keywords, doing so would deprive me of an opportunity to report yet another spam source to the appropriate blacklist. Also, I don't like the idea of limiting the vocabulary that can be used in e-mail to words that no spammer would ever send. I'd rather blacklist 99 percent of the entire Internet address space than even one percent of my own language.