[Mailman-Developers] Mea Culpa.

The guy named after an Om Kalthoum song ocschwar@MIT.EDU
Sat, 11 Nov 2000 00:40:44 EST


When I first set out on this I was already
planning on setting up a server with an
encrypted mailing list in order to invite some
friends of mine to try out the social dimension 
of it.

So, I quickly concocted a gross misunderstanding in
my mind about how mail transfer agents and mailing list
managers divide up their duties. Thank you for 
correcting me on that.

I also was a little optimistic on the idea that
crypto email fora being the final element to
establish pervasive crypto. A look at cypherpunks
shows I'm dead wrong. What I want to do was attempted
in the past and the code just died from obscurity.

But, let's say I still set out on this:

On the MTA side (I'd probably diff Postfix) I would 
have to enable the MTA to know to divert mail coming
in to certain lists over to a side spool, activate the 
crypto-exploder, and then spool it to outgoing. Then comes
writing the crypto-exploder, which would be a simple
Perl or Python script invoking relevant the GPG and MTA modules.

On the MLM side, all that really is necessary is for Mailman
to be able to collect and revoke public keys(/etc/pki?), 
and deliver its own public key to those who request it.
A host-owned (rather than user-owned) key ring has been discouraged
in theory, since it would prolong the life-span of a revoked
public key. Any server that used one would need to check
in with a keyserver on a cronly basis.

Regardless of the MTA issue, a GPG-enabled Mailman
would be convenient. You could automatically process
signed transaction request emails, and have the
admin manually process unsigned ones, for example. 
So if I do it I hope you'll accept the plug in.

Regards,

-- 
Omri Schwarz

 I get serious letters from university students, asking questions for a 
 project they are doing - these are not much different from those I get
 from school-children (written in green crayon), except the writing is
 a little worse. -- Terry Pratchett, Warwick Uni (10.11.94)