On 2018-01-09 10:25 AM, Bill Cole wrote: > A better place for this discussion would be the MailOps list, where a > broader variety of mail admins *INCLUDING MS EMPLOYEES* take part, and > this problem class has been discussed multiple times. See > https://chilli.nosignal.org/cgi-bin/mailman/listinfo/mailop
Thanks for the pointer. I may join when I will have time to engage. > With that said, a couple of paragraphs stand out as demanding response. > They literally made me choke on my coffee, repeatedly: Sorry for the coffee, I hope you did not get scolded, stained, or keyboard spilled. Happy to offer you a replacement coffee over a reasoned discussion of the issues. > This is why operations like DocuSign & Greenvelope > exist. Email is a terrible medium for exchange of legal docs, but that's > a tough sale to lawyers... I am a lawyer and I am all ears. I am listening to my clients and using the medium they choose to communicate. I've seen some DocuSign use here and there. Not sure how is that better than email since after all it is an email with a link to click, so the whole issue of legitimacy, authenticity, etc. is just displaced. In the end, all we need is a single-purpose tool to reliably send a sequence of ones and zeroes between the different parties. Wether that tool is HTTPS or SMTP+TLS or some other form of electronic signaling, I am agnostic. >> The only minor objectionable issue I find with my email is >> the GPG signatures: I sign with a key that is not associated with any >> email address, contrary to RFC_dont_remember_which, > > And totally against the whole design of the OpenPGP signing protocol. > You might as well use a key claiming to be from someone else. This is > WORSE than not signing at all. It screams "THIS EMAIL IS FRAUDULENT!!!" The OpenPGP signing protocol that has miserably failed achieving any level of significance outside of specialized IT engineering circles for twenty years? The protocol that mixes up identity with address with authentication with authorization? Happy to continue this discussion in the appropriate forum. Open to be persuaded that for a decade RFC4880 has been hiding the solution to the faults of RFC1991 and RFC2440. I remember reading them back then, but I do not have my notes of the time. Bottom line: I live at an address, I am not that address. I can live at multiple addresses, with multiple people, and the way the OpenPGP signing protocol connects keys to email addresses is a bad abstraction with many bad consequences. >> and I sign with an expired key > > So it looks like you're using a key that you have in the past said that > you wouldn't use at this date. This is practically begging to be > distrusted. Separate content from transport. In analogy to snail mail: separate letter from envelope. To trust or distrust the content is a job for the recipient or its post-delivery filter, not for the MTA. Input from the post-delivery filter or post-delivery user interaction may feed back and inform the receiving MTA's future rejections (i.e. the reputation of the sending MTA). An MTA should only accept or reject. If it accepts, it must deliver. Once delivered, open the envelope and analyze the letter. That's where you can decide to trust or distrust. Sort into inbox or junkbox. If the recipient wants to risk of losing messages, feel free to drop them. However, not silently drop them. Leave a message, a textual notice to the recipient with the empty envelope. Fax machines are still wildly popular because there is a proper protocol in place and when the sender receives an OK, it can rely on the certainty that the message has been received. If the recipient's dog eats the paper on the other end of the transmission, it is a liability for the recipient, not for the sender. >> (my bad: good enough for my purpose and therefore very low >> priority to fix). > > This free-floating assertion seems at odds with the fact that your email > to Microsoft customers is being treated in a manner that only makes > sense for phishing or malware email. I can explain the appearance if you are interested. No contradiction. > Have you tried taking the 5 minutes required to set up a correct GPG key > and use that instead? As you can see at <http://pgp.mit.edu/pks/lookup?search=Yuval+Levy&op=index>, not in a long time. Because if you look at <http://pgp.mit.edu/pks/lookup?search=Bill+Cole&op=index>, you see how spambots can harvest email addresses from key servers. Yay, who needs namespace mining? Plus, the resulting list is very targeted: highly intelligent people who take the time to set up GPG keys. A spammer's dream. Some other interesting information can be gleaned that may not be intended to be public. I live at an address, I have multiple addresses, I do not need the public to know any of them nor to link all of them with my person/identity/activity. The PGP protocol is almost as bad as using a fingerprint sensor, or facial recognition, to unlock your phone. > It is not a bad idea to reject email with a general format that is > widely used by scammers, discussing activities often discussed in such > scams, bearing a doubly bogus "signature." > > In fact, that strikes me as a very GOOD idea. Continuing your line of thoughts, maybe it is an equally good idea to reject SMTP because it is so widely used by scammers? If I look at my server's statistics, over 90% of messages are rejected (thanks Spamhaus). Let's apply Pareto-efficiency and get rid of the remaining 10%. Scammers will inevitably mimic the legitimate protocols and the legitimate formats. The way to deal with them is not make the legitimate protocols and formats unusable. The way to deal with them is to increase their cost of participation to the federated protocols. Easier said than done, I admit. >> Coming from the company that has gifted the world the scourge >> that is HTML email, > > Niggle: Outlook Express added HTML email in response to that > "innovation" in Netscape Navigator Gold 3.0. Have you ever observed the adoption of new ideas? Think committees at work. The first mover takes a risk, and it depends on whether the move is seconded or not that the first mover is then seen as the fool that came up with the rejected idea, or the visionary that came up with the next big thing. Similar in everything. The mouse may have been invented at Xerox, but its adoption was driven by others -- and even though it was seconded by Apple, I'd argue that it is Microsoft that made it truly popular. Yuv
