Re: [dtn] draft-blanchet-dtn-email-over-bp-00.txt

[Stephen Farrell <[e-mail]>]

Hiya,

On 17/01/2023 23:37, Marc Blanchet wrote:
> Cool. But a drawback to me is that it seems that the solution you are
> describing is a bit implementation specific. I was trying pretty hard
> to stick to protocols, so it is more clean and interoperable. Our
> goal, as a community, now is to standardize stuff to achieve a high
> level of interoperability.  And includes the assignment of a service
> number so all implementations expect the same payload at the same
> place...

Yeah, but.... I'd argue there really are a pile of reasons to
prefer an architecture where one uses the DTN to sync message
stores over one where one tries to do SMTP/BP. Just to pick
two: there was no chance I could convince our email admins to
run any BP stuff, but syncing message stores doesn't require
that permission, and secondly, I could configure which mails
from my normal email a/c to forward into the DTN. That kind
of discussion is better f2f with whiteboard of course so may
be one for an upcoming IETF meeting.

Yes, what we did was implementation specific, but I'd bet one
could do something with SIEVE or IMAP that was standardised.
(That said, our implementation was really to commit all the
files in each to-by-sync'd Maildir folder to a mercurial repo
(IIRC) and to then encapsulate the ``hg diff``'s in bundles
so yeah, there was a big element of just doing what worked
but the diff thing was significantly better and e.g. even
handled cases where a mail was moved between folders as well
as read/forwarded flags, which are important.)

Architecturally, I guess one could start by asking if one is
trying to make SMTP or mailboxes available to users who are
in challenged regions of the DTN. I guess I'd argue the
latter is far more useful.

> Having said that, I would love to get the info. Have you done an
> internet-draft describing it.

Sorry, no. I can probably try dig up some paper but forget
how much implementation detail we documented. Will look in
the morning (here:-)

Cheers,
S.