On Fri, Sep 27, 2019 at 7:15 AM Dave Crocker <dhc(_at_)dcrocker(_dot_)net>
wrote:
On 9/27/2019 1:43 AM, David MacQuigg wrote:
I remember learning about email systems, primarily from RFC 5598. I
found Section 2 frustrating and had to give up and develop my own model
at the administrative level. I have found this model very useful in
teaching and in developing my own email system, box67.com
<http://box67.com/>.
1. If you only found that one section frustrating, you are doing better
than most others, including me.(*)
Compared to other RFCs, I would say 5598 is excellent.
2. The thing about RFC 5598 is that it was developed in a consensus
process. So whatever its faults, it represents a meaningful degree of
community compromise. For this topic, no other equivalent document can
say that.
Agree.
3. Pursuing discussion of improvements to RFC 5598 would be entirely
reasonable.
Wow!! My first thought is "let's jump right in". But then I think - no, I
don't have time for this. I haven't been working with email systems for
over ten years, and my main focus right now is managing a horse ranch.
4. (This adapts from my original note) Using language that has no
shared understanding prevents communication. If one wants to make a
point with other folks, one needs to use language that is well
understood by the target audience.
Perhaps we could view an elaboration of the Administrative Level as an
extension of what is in RFC 5598. We need better definition of the
specific roles of the Actors, not just call them all ADMDs. We need
something equivalent to Fig. 1 in
http://en.citizendium.org/wiki/Email_system
5. Your text, quoted above, cites an alternative bit of work, but does
not provide a citation to it. Instead there is a generic citation to
work that 'uses' it, apparently. Following that link produces a page
that starts with "Much of what is described in these pages is not yet
implemented." So to the extent you meant to give us something superior
or even useful, that isn't the way to do it.
See the Citizendium link above for a general introduction. See the links
at box67.com for details on the system I put together. This is a
whitelisting system based on the reputation of the Transmitter's domain
name. It uses DNS to host a database of authentication and reputation
information. It worked even better than I expected. Work stopped in 2008
when I started using gmail, and was no longer bothered by spam. I haven't
kept up with developments in the last ten years, but I suspect gmail and
all the other services that are very good at blocking spam, are using some
form of proprietary reputation system. That's a shame, because an
open-source system would have the possibility of becoming a de-facto
standard, and that would get us to a zero-spam mail system (at least for
those who would care enough to follow the standard).
_______________________________________________
ietf-smtp mailing list
ietf-smtp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf-smtp