Living at an Earn site, therefore representing a minor form of life,
I'd like to ask somebody at the bright side of this network to support
the following suggestion (half a :-) here)
 
Furthermore being a rebel running the Bitnet version of Listserv
I'd nevertheless like to ask Turgut to consider this suggestion
(the other half of a :-) comes here)
 
An increasing number of sites is using some flavor of the Netnews
system under VM, VMS, and Unix to read Listserv mailing lists.
Thanks to Eric who supported this Listserv-list-to-Netnews-groups
gatewaying by adding the Newsgroups definition, and more important
by creating a message-id if necessary. I think this support is
worth mentioning as the origins of Netnews are Unix based, and Eric is
a Unix hater (or at least was at that time).
 
Now the message-id will only be preserved for FULLHDR subscriptions.
Even if you know about this, the situation is sometimes somewhat tricky.
If a gateway subscribes to a non-peered redistribution list that is
subscribed with SHORTHDR to the main list, the original message-id
is lost. Usenet is based on redundant feeds, most of the suppression of
duplicates is done based on message-ids.
 
There may be more systems out in netland taking advantage of message-ids,
but I'm not familiar with them, and don't know if it's a nice extra
or a must.
 
In a perfect world each gateway is subscribed to the right list using
the right options. Now look around, do you see a perfect world? (If
your answer is yes, please contact me offline)
 
You guessed it, my suggestion is to preserve the Message-id for all
forms of header options.
 
Disadvantage:
- One more header line for most subscribers.
 
Advantages:
- Avoid risks for unexperienced (with respect to Listserv internals)
  gateway maintainers.
- Maybe an overall traffic reduction (I'm not 100% sure here, and too
  lazy to check it out). Some subscribers (and probably an increasing
  number of subscribers) are setting FULLHDR just to get the message-id.
  This causes two distribute jobs where one would be sufficient if
  the message-id is preserved in SHORTHDR mailings.
 
Thanx for listening to my babbling,
  Thomas