joe@hanauma (Joe Dellinger) (09/04/88)
Our machine, "hanauma", is connected to a few sites (notably decvax for historical reasons) via uucp links and is also on the arpa net as "hanauma.stanford.edu". When I send mail to some distant site, usually it is best to get to the local area via the arpanet and then make some uucp hops to get where I want to go, ie "isis!timna!seb@boulder.colorado.edu". The problem comes when somebody tries to invert this route. Machines with "smart" mailers (such as isis) grab the mail destined for us, no matter what the specified routing, and send it on to us through decvax, which can take a week! I had thought that the correct way to announce that you were also on the arpanet was to list yourself as a possible gateway; this turned out to be an incredibly bad idea as machines all over the country decided we were the BEST Stanford gateway, and started routing all Stanford mail through New Hampshire! Is there something I'm missing here? I'm trying to collect the e-mail addresses of as many Geophysicists interested in Anisotropy as possible, and so people have been sending me "their address". So far, most of the addresses I have received have been arpa-uucp mixtures, and most of them have failed at the ARPA-uucp junction when I've tried them. Are things really this bad, in general? Or is it just that Geophysicists make rotten computer scientists? (Comes from being taught that FORTRAN is a computer language, perhaps.) \ /\ /\ /\/\/\/\/\/\/\.-.-.-.-.......___________ \ / \ / \ /Dept of Geophysics, Stanford University \/\/\.-.-....___ \/ \/ \/Joe Dellinger joe@hanauma.stanford.edu decvax!hanauma!joe\/\.-._