>Looks like we really have two DIFFERENT backbones... The
>problem of compatability between the two is getting to be
>really annoying.
 
Yup, let's split the backbone. Even if somebody is working
on a gateway, at the time being there is no such beast.
Nevertheless, let's split the backbone. Don't look for
any consequences. It will either increase the load on
some already busy lines, or burst the last-hop-before-the-
other-side-of-the-fence Listserv, or both. Nevermind, lets split
this useless backbone.
>
>Is it possible to treat it as two DIFFERENT backbones? Each
>having it's own PEERS NAMES. And having a (multiple?)
>gateway between the two which could resolve the differences.
 
What gateway?
 
But why stop here? I'm now in the right splitting mood,
well, it's more a splitting fever. Let's separate those VM rel 5
(or is it HPO?) sites that still messes around with spool
files. They may form their own backbone. And then there are
some sites which cause trouble because they have not updated
Bitearn Nodes and their Mailer tables for years. Of course
they should have their own backbone.
 
Why stop at Peers Names? I'm tired to read each and every
month why Bitnic is not able to provide the node table
in time. Let's nuke this entire network, it would save
us all a lot of trouble. Every node forms its own network.
No trouble, no hassle. Most of you probably remember the
discussion about the bandwith of a station waggon full of
tapes. Technology improved in the meantime, it will be hard
to beat UPS and Exabyte tapes anyway.
 
Steve, this no personal flame (if it were, I had sent it
as personal mail). I know that changing a node name
is trouble enough even without additional misrouted Listserv
jobs. But I would prefer discussing how to improve the
situation without using an axe.
 
I guess some Listserv maintainers were not aware that some
magic refreshed their Peers Names in the past, and that
this magic stopped working in Europe. And if they were aware,
they didn't realize the consequences. Well, now they
should know, and PLEASE immediately refresh their Peers
Names file, either by fetching it from the TREARN server,
or from wherever they can get an up-to-date copy.
Don't think by looking at your mailing lists that you
are not affected. A user may send a distribute job, that
either must be processed manually or goes to a bit bucket if an
out of date Peers Names is used to construct the distribute
job.
 
  Thomas
 
P.S.: Regarding Splitting Image: Spitting Image is a
satiric TV series in the UK. Dunno why this network reminds
me on satiric TV shows ...