On Thu, 11 Feb 1993 16:42:57 CST David E Boyes <[log in to unmask]> said:
 
>I have a version that shows 1.7d in the exec comments in production. The
>version available  from LISTSERV@SEARN as  of today still  shows version
>1.5o in LSV$FWDC. The other exec (LSV$FORW) shows version 1.7e.
 
Yes, FWDC (the code that handles  the /FORWARD command, as opposed to the
actual forwarding) hasn't changed in ages.
 
>I'm  not  sure  I  agree  with  this  statement.  Given  that  NOTE  and
>PROFS-format mail are  still officially acceptable (*I*  don't think so,
>but the  NIC still won't declare  either to be  bogus) and a lot  of Vax
>sites still  don't have or  use a mailer, people  end up with  mail sent
>directly to their  reader. Linking the two methods  together generates a
>consistent  behavior --  anything that  looks like  mail gets  forwarded
>properly.  The  inconsistency of  only  having  certain items  forwarded
>because of stupid configurations at other  sites is a serious problem to
>the end user.
 
I still  stand by my statement:  having the two methods  available to the
user is not  a good idea, precisely because of  the inconsistency it will
generate if  the files  are not kept  in synch. The  problem is  the same
whether you  use XMAILER+LISTSERV or  LMail+LISTSERV: you have to  find a
mechanism to keep the two in synch, run only one, or expect problems. One
possible  solution is  a documented  EXEC  that issues  commands to  both
servers, if  your users are docile  enough. Another is an  EXEC you'd run
once a day in one of the two servers to update its forwarding information
based  on that  of the  other, which  would be  the "master"  information
source.
 
I  can't  think of  a  good  synchronization  mechanism between  the  two
servers, given  the difference  in functionality. That  is, LMail  can do
more than LISTSERV in terms of forwarding syntax and support for multiple
forwarding addresses,  so it  wouldn't work to  have LMail  tell LISTSERV
"Hey, Joe  told me  to do that,  so you make  the same  change". LISTSERV
could give  such orders  to LMail but  what is to  prevent the  user from
issuing commands directly to LMail? A synchronization mechanism must work
both ways. And then there is  the problem of statistics, which is another
can of worms. If it were a 50 lines change, I'd be happy to do it.
 
  Eric