dhesi@bsu-cs.UUCP (Rahul Dhesi) (10/06/88)
In article <Oct.5.23.13.39.1988.12866@NET.BIO.NET> lear@NET.BIO.NET (Eliot Lear) writes: >There are essentially three benefits of using real domains over .UUCP: ... >2] It increases the name space. There can only be one foo.UUCP, but > there can be a foo.rutgers.edu and a foo.berkeley.edu, etc. This is probably true for now. (Wishful thinking follows.) It doesn't have to be. I can imagine a modified domain system, which increases the namespace without eliminating the UUCP domain, in which: foo.berkeley.UUCP is a valid domain address (and "valid" is not synonymous with "registered"), mail to foo.berkeley.UUCP goes to a server for the UUCP domain that then relays it to berkeley.UUCP, that relays it to site foo.berkeley.UUCP, whose whereabouts might be known only to berkeley.UUCP. Only a few sites need be at the topmost level within the UUCP domain. Initially all UUCP sites will be something.UUCP but this number would decrease as more of them hid themselves from the top level and relied on relay sites. The current UUCP topology is fairly tree-like already. -- Rahul Dhesi UUCP: <backbones>!{iuvax,pur-ee}!bsu-cs!dhesi
jordan@zooks.ads.com (Jordan Hayes) (10/08/88)
Rahul "I have one physical transport mechanism on my machine, so everyone
else must be set up the same way" Dhesi <dhesi@bsu-cs.UUCP> writes:
[ a bunch of cruft that basically outlines the DNS and has
but one extra "feature" -- it includes a top-level "domain"
called "UUCP" -- presumably so he doesn't have to fix his
dead mailer, and surely one in which his so-called "domain
server (read `relay') for UUCP" wouldn't be *his* machine ... ]
Rahul, aren't you forgetting something? Come on, you can remember,
think back ... there's something you're forgetting ... remember?
RIGHT!!! Duh, it was so simple, so close to your grasp!
"Naming does not imply routing"
/jordan