Here is a real world example of the cost of header style proliferation!
I have a list whose owner just couldn't make up his mind which header
he wanted for the default.  So over time he changed it and changed it.
Now when there is a post to that list LISTSERV produces
 
Done - 14 jobs (75 rcpts), 2 outbound files (317 rcpts).
 
For the folks subscribed with SHORTHDR.  The reason that there
are two outbound files is that there is also one subscriber
with SHORT822 getting direct delivery.  Then I get:
 
Done - 6 jobs (27 rcpts), 1 outbound file (88 rcpts).
 
These are the people who subscribed when the default header was
FULLHDR.  Most of those distribute jobs are going to the same
hosts that got the SHORTHDR distribute jobs.  So it is not only
more work for my system but also for quite a few others.  Then
comes:
 
Done - 1 job (4 rcpts), 1 outbound file (18 rcpts).
 
And these are for the folks who subscribed when the default was
IETFHDR.  The only header style I'm missing is FULL822 although
there may be some in those distribute jobs.  I really, really, really,
do NOT want another header style.  No thank you very much!
 
Eric is being very generous in allowing 5 minutes to correct this
in sendmail.  How long do you think it would take to edit sendmail.cf,
insert the line
 
Oh64
 
or if you are using v8.7.x
 
O MaxHopCount=64
 
and restart the sendmail daemon.  30 seconds is more like it.
 
Just yesterday I deleted a DIGEST subscriber in Australia from one
of our lists.  Digests are sent with what amounts to the SHORTHDR
and that subscriber was more than 17 hops away with an old sendmail
in .AU bouncing the digest because of the hop count.  A SHORTHDR
with MIME data in it is *not* going to solve the problem.
 
-----
Glenn A. Malling <[log in to unmask]>
Syracuse University Computing Services   +1 (315) 443-4111
220 Machinery Hall  Syracuse, New York 13244-1260
 
Postmaster for SUnix and LISTSERV.