>>     Also, maybe KERMIT
>>     could be distributed out to several listserv sites?
>I'll let the list answer this one. Not me -- not enough DASD space.
 
I'm glad that something is going on with the Kermit distribution.
I have offered two times within the last year(s) to help in reducing
the transatlantic distribution problems for Kermit files.
 
The crucial point is that something like a distributed caching
file server system would be needed to guarantee the necessary
user transparency. We don't have enough disk space to hold the full set
of Kermit files (and I have severe doubts that there is a site
anywhere which has that amount of free disk space). So we could
only store the most important variants (something like the
versions for IBM-PC, VAX, C-Kermit, and maybe some more).
Requests which may not be satisfied by a particular server
must be forwarded automatically to the main backbone. I assume
this could be realized by means of Listserv's exits.
 
Furthermore it would make a lot of sense if the main backbone
server (KERMSRV@CUVMA) could forward requests to another server
that is nearer to the requestor (and has the requested file available).
But I have no idea how this could solved with the existing software.
 
Anyway, I'm ready to give some disk space and CPU cycles to this
project. What I can't offer is time to write even a single line
of software. I assume that installation and normal operation
will eat some time too, but that's ok if it lowers the network traffic.
 
Hoping that it will be easier than I believe to
set up such a distributed server system
 
  Thomas