On Tue, 19 Sep 89 17:59:13 GMT Eric Thomas said:
>                   However, the introduction  of this new line caused LISTSERV
>distribute jobs  to take  completely "absurd"  paths, because  I had  not been
>warned about  this and had  therefore not made  the appropriate change  to the
>link weights file; LISTSERV,  which has no way to know that  this is a virtual
>SNA route, had assumed  the default weight of "1" and routed  a huge amount of
>traffic through that line.
.....
>I am hereby  repeating, for the last  time, my request to have  changes to the
>international EARN topology announced a  reasonable amount of time in advance.
>I am fed up with having to find out this kind of things after the fact, when I
>get complaints from understandably unhappy  people.
.....
 
Eric, you've complained for as long as I can remember that you simply
have no way to fully understand the topology of the network, particularly
keeping up with all of the changes like the one you describe above.  Yet,
you are the only person capable of generating the LINKSWT files?  Is this
not the time to consider letting someone at BITNIC
   (Side point: I presume that you're only generating this file for
    non-EARN sites, right?  And that Turgut is doing it for the EARN
    sites?  That's why I'm suggesting BITNIC here.)
take over the generation and maintenance of the LINKSWT file.  It would
seem that since they're aware of the changes in the topology faster
than most people, since they're getting the data together for the
rest of the network definition tables we use, that they would be a
reasonable place to handle this chore.  Also, given they have a
more network-wide viewpoint, they may have a better understanding
of the topology than one person.
 
This is not a flame against you, Eric.  I'm merely trying to suggest
an appropriate person to take over something you've called onerous in
the past.
 
Richard