Possibly, I will defer to the more technically learnt.
   I'm not a nym server expert but from my laymen perspective the Pynchon
   Gate design looks good. It might be totally redundant and unnecessary
   but if metadata analysis is the concern, wouldn't such a setup be even
   more secure by coding something so that the time between sending a
   message and receiving a reply which in theory could leak information
   about the nym holder, be sent at a random date in a given time-frame
   (unbeknownst to the metadata leeches) . i.e. In 6-12 hours from the
   moment I click "send" or say in 12-20 days etc.
   The email message could be coded to send at random like an online
   roulette table ball, within a given time window: verses say reloading
   every 24hours. This would in theory give out incorrect message 'sent'
   time-stamps, or would this be unnecessary because traffic from the user
   to the email distributors is already being controlled by the user,
   which queries into intervals anyway? Is that not metadata that can be
   tracked?
   - J
   On Wed, Aug 28, 2013 at 12:22 AM, danimoth <[1]danimoth@cryptolab.net>
   wrote:

     On 27/08/13 at 10:22pm, Jeff Scofield wrote:
     [cut]

   > One strategy might be to consider the adoption of a time delayed
   email
   > system.  The reason why the use of such a mechanism to allow someone
   the
   > ability to write an email, and then have it sent off at a specified
   (or
   > randomly generated unspecified) date is useful for multiple reasons.

     [cut]
     Are we trying to reinvent anonymous remailers and nym servers?

References

   1. mailto:danimoth@cryptolab.net