John is right about the problem.
SMTP email has become a cartel, if you are not in the club, you can't send email and there is no way to fix that within the SMTP protocol because there is no way to express authorization.
POTS service has the same exact problem, the telephone system is dying because it is clogged up with fraud and spam. And we can't expect corporations to fix this either, Facebook can't be bothered to address the disinformation spam masquerading as 'satire' and it has a closed platform.
The only way to effectively address abuse in an open service is to apply authentication and authorization to every post where the authorization criteria are chosen by the recipient.
One way we could do that for IETF is to shift from mailing lists to NNTP or IMAP. Both of which guarantee that the recipients can get the mail they have agreed to receive because there is an opt-in. Problem there is that the usability of such schemes is really poor because it isn't something the email clients are set up for. Which is why there is limited uptake for the services people have put together for this from time to time.
Another way would be to switch to a Web forum reader that had adequate affordances for users. But that also ends up being kinda sucky because IETF is separate from my main workflow. I stop following groups that move to using GitHub for communications because I just don't follow those, they are separate from my main workflow. Also, I find the PR mechanism to be incomprehensible. If I want to make a comment, I want to make a comment, not fork the repository, make changes and ask for the changes to be approved. Some people can wrap their minds around that, the diminished engagement in such groups suggests others can't.
The devil is of course in the deployment. Proposing a better way for 2000 people to work is easy, getting the transition is hard.
Now that I have the GUI version of the Mesh client about to release, I am turning to the problem of collaboration. I already have some assets here:
1) An engine that allows posts to be added to an append only list and distributed.
2) An engine that will parse Word, Markdown or XML2RFC format and produce Internet Draft format in any of the input formats or TXT.
3) A WebView client that runs on Windows, Mac, Android and iOS.
4) A TKI that manages keys to effect end to end security for all of the above and provides an end-to-end secure asynchronous messaging layer that can be used to initiate synchronous communications[*].
The Mesh is designed to enable end-to-end encrypted social media but I don't think we need to go quite that far for this particular application. What I plan to build is a simple comment on Internet Drafts mechanism that can be interacted with through a Web front end, a gui client or email.
Now imagine that people start using a tool that integrates into their workflow, letting them know that there are things to see in the places they look. And that tool supports all the messaging required to support a Working Group.
Basically MeetEcho but for Working Group process. So when people mark up a document, they can differentiate a nit from a security issue or a new requirement or identifying an error.
[*] I was looking at WebRTC but MOQ is the better choice.
On Sat, Jun 29, 2024 at 12:57 PM John Levine <johnl@xxxxxxxxx> wrote:
It appears that Salz, Rich <rsalz@xxxxxxxxxx> said:
>
>>I do not see why the IETF needs to make use of a "major" email-delivery
>>service at all. Surely there are some mid-tier or similar equivalents.
>
>Because these days if you're not major, you are more likely to have your mail blocked or treated as spam.
That's basically it. Since 90% or more of all email is spam, it is
entirely reasonable to assume that mail from a source you haven't seen
before is spam. (This happened to me when I renumbered my IPv4 network
last month.) The more mail you want to send, the harder it is to get a
good reputation.
It is my perhaps naive belief that most IETFers would prefer to get
their mail over IPv4 than not to get it at all. We're looking at IPv6
options and I hope we will find one that works, but I can tell you
from experience that it is not trivial, and it is not quick.
R's,
John
