MACALLSTR@vax1.physics.oxford.ac.UK (07/23/87)
It makes me VERY CROSS to hear people criticising JANET addressing especially
when the schemes they're using are definitely inferior.
On JANET, based on 'X25' packet switching, I can address ANY SITE simply by
specifying
username @ sitename
I don't have to work out the route my message will take. That is taken care
of by the 'network'.
Furthermore, on my VAX/VMS system on JANET, I can address ANY SITE with a
GATEWAY to JANET simply by specifying
network%sitename::username
This gives access to any site in the WORLD with a very simple addressing
scheme.
On a network such as JANET every site uses the same network address for
a particular site and the user is insulated from considerations of routing
at a very low level. This cannot be said of networks which use routing
from one node to the next in a very complicated and very unreliable structure.
I'd suggest that the problems in sending messages to JANET are nothing to do
with JANET but arise because the intermediate routing nodes have not defined
the appropriate names - a critical weakness of distributed routing schemes of
this nature.
John
P.S. No comments about JANET domain ordering PLEASE!