[ih] ih ML-manager configuration strips signatures(++)

[Dave Crocker]

On 7/10/2020 7:29 AM, Steffen Nurpmeso wrote:
> Dave Crocker via Internet-history wrote in
> <c4f8c29f-5414-00aa-cf3f-f0441f6fa64e at dcrocker.net>:
>   |On 7/9/2020 6:41 AM, Steffen Nurpmeso via Internet-history wrote:
>   |> (A little bit
>   |> funny that now that everybody goes for DKIM and we see this
>   |> terrible "Name via List <list>" everywhere, personal and conscious
>   |> signatures end up mutilated.)
>   |
>   |fwiw, that's an artifact of DMARC, not DKIM.
> 
> Oh yes, sorry, not meant to offend you. 

No need to apologize.  People mix the references all the time.  I am 
picky about it in technical venues, to make sure people can focus on 
considering the precise technology at issue.


> But still, whereas
> i really see DKIM, i do not like the way it is used.  I have seen
> messages where several intermediate receivers each performed DKIM
> verification, which i could possibly understand / deem ok, but
> also generating DKIM anew.  Isn't that a tremendous waste of
> resources of all kind, need- and useless, and how it increases
> header size.  Just imagine every party along a traceroute chain
> would reensure the origin (if it could).  No no, original sender
> (create), a possible mailing-list manager on top (verify
> + create), and final receiver (verify), that would make sense to
> me.  (The problem with the OpenPGP but especially S/MIME i like
> more is that delayed authentication may no longer be possible.
> And that it needs MIME.  And that the huge graphical / web mail
> applications may not handle it nicely.)

Yes, it is inefficient.  No, it's not a problem.

(If we wanted to make email or the web highly optimized, there are many 
things that would need to be done differently.


d/
-- 
Dave Crocker
Brandenburg InternetWorking
bbiw.net