Before the $'s going down farther 2 cents:
Although LISTSERV  is becoming  a de-facto standard  it's not  an "official
network tool".  There exists a "coordinating  board" (3 ppl. ?)  but except
for  the INFOs,  PEERS  NAMES and  LINKSWT stuff  there  is no  centralized
concept in it.
This doesn't mean that the BITNIC  LISTSERV was more "official" or suitable
for automated update.
*-*
LISTSERV can be seen from a "local" or a "network" point of view.
I can use LISTSERV  (with or without things like MAILER) as  an easy way to
support lists  without worrying  about DISTRIBUTE and  other things.  I can
install LISTSERV as a  fileserver. I can modify it in  whatever way I like.
Or:
LISTSERV can be used to reduce  traffic significantly. LISTSERV can help to
get a more fair consumption of resources and many other things.
*****************
Just read Eric's note - will we lose him or not ?
Eric: don't leave before 1.7q !!!
      and - I hope you get paid in FF not in $ !
*****************
*-*
If we make an "official LISTSERV-subnet" I would vote for:
 1a) centralized and automatic updates
 1b) unmodified code  - i.e. sites with "local mods"  must either guarantee
     that those do not affect the function or can't take part
 1c) strict rules for "LISTMASTERS"
       >>>>>>> OR <<<<<<<<
 2) a "stabilzed"  Version of LISTSERV, no updates within  a certain period
     of time. I.e. no modifications even if some feature seems desirable.
This would put more weight on Eric's  shoulders. :-) - I just can't rely on
the OPERATOR  to re-start LISTSERV in  case of an "update-crash".  AND this
must also include  some circumvention/bypass in case one of  the servers is
for whatever  reason out-of-service if  the proper update is  important for
the function as with the 1.5x sub-releases.
*-*
Although I  would accept "network-update" of  the server I am  against this
feature if not "all" will allow it. That means in detail:
 a) all "official" LISTSERVs run at the  "same" level and all the update is
    done centralized. It is also assured that:
    1) (immediately) pre- and post-update LISTSERVs can work togehter
    2) the  next  update  can  only  be started  if  the  previous one  has
       completed succesfully
 b) "unofficial" LISTSERVs are neither  part of the DISTRIBUTE-backbone nor
    can they  act as Peers. If  they are hierachically linked  to some list
    they must behave like a "general user" (whatever this means :-) )
 c) not  precautions are  taken to guarantee  the finction  of "unofficial"
    servers
 d) the responsibility for UOS is fully by the LISTMASTERs
*-*
Some final remarks:
* I experience the mess with the  MAILER tables (which is not Alan's fault)
* I can remember  how long it took  until our node was "known"  to at least
  90% of the net
* and I see  how lomg it takes for "standards"  to become standard (RFCxxx,
  the POSTMAST(er) ID on  every site you can send to, the  INFO ID at least
  in EARN, ....)
*-----*
I dunno if it's worth the 20m$s ...  but anyway I will see the response not
before April 27th when I will be back from vacation.
Christian