jsdy@hadron.UUCP (Joseph S. D. Yao) (11/27/85)
I am impressed. This is quite an effort; public thanks to steve wall aka snoopy. One caveat I think must be added (since I fell into it >1ce in younger days). In article <9072@ritcv.UUCP> spw2562@ritcv.UUCP (snoopy) writes: >UUCP ==> Internet hosta!...!hostz!seismo!user@host.ARPA or > hosta!...!hostz!harvard!user@host.ARPA > Note: "hosta" is the sender's nearest neighbor. > The sender must specify a UUCP routing path to > "hostz", a host with a direct connection to some > UUCP->Internet gateway, such as "seismo" or > "harvard". None of hostb ... hostz may be on the ARPAnet or some other net that will swallow @'d addresses. For instance, if I want to send ARPA mail from aplvax and forget that umcp-cs is also an ARPA host, I might construct: umcp-cs!seismo!eric@ucbvax.ARPA Well, when it gets to umcp-cs, the address is: seismo!eric@ucbvax.ARPA which gets shipped across the ARPAnet to ucbvax, which then looks for seismo!eric Well, this is a bad example since ucbvax just happens to call seismo right back, cross-country (3250 miles), and I think that there is an eric there -- but not the one i was sending mail to, on ucbvax! [Potential flamers -- I didn't say HOW ucbvax called seismo back.] The bottom line is, that you want the shortest UUCP route to an ARPA server. You may have gathered from above that it's possible to go through the ARPA net and then back onto UUCP, which is true, and left as an exercise for those who don't want to send a stamped self-addressed packet envelope. Incidentally, seismo.CSS.GOV (sometimes incorrectly identified as seismo.arpa) has been complaining that so much mail passes through their site every day, that if they charged a penny a kilobyte they could all retire as billionaires. One report says that factions are split between doing that and losing their mail-forwarding software. ;-) Seriously, if anyone can find a site other than seismo to forward through, I know some folk who'd appreciate it. -- Joe Yao hadron!jsdy@seismo.{CSS.GOV,ARPA,UUCP}