vixie@decwrl.dec.com (Paul Vixie) (10/27/88)
[ This comes to me from Jordan K. Hubbard who now works for PCS Computers in West Germany. His network address is unido!pcsbst!jkh@uunet.uu.net. I want to draw your attention to the Return-Path: and From: lines on the mail msg as it came to me. So, here we are with somebody else's opinions on "active rerouting". Jordan can't post right now due to some kind of technical problem. Note that I do not really agree with Jordan's arguments -- active rerouting would be bad even if the maps were perfect. But you all know what I think, here's what Jordan thinks. --vix ] Return-Path: unido!pcsbst!meepmeep!jkh@uunet.UU.NET Subject: Re: ARGGHHHHHHHHHHhhhhhhhhhh..... Date: Wed, 26 Oct 88 15:32:59 N From: unido!jkh@uunet.UU.NET (Jordan K. Hubbard) >Jordan K. Hubbard writes: >> Sigh. Don't tell me, let me guess. unido is an *active* re-routing >> site, yes? > >unido!postmaster replies: > >True, and let me say it right away: We won't change that!!. > >> -> unido!host!..host!user >> >> No no no no no no nononononononono.. Leave it be. Hands off. No touchie.. >> > That is exactly the case we want to catch! There are enough people > giving us silly routs like routing from Germany via USA to Germany > and so on. Oh dear. I was afraid of this. I have absolutely no desire to replicate the debate currently raging in comp.mail.uucp between the likes of Eliot Lear and Chip Salzenberg (thought the postmasters at unido might do well do read it, even Eliot (the pro-active rerouting person) admits to the many flaws inherent in it). However, I still feel very strongly that re-routing is a very bad, shortsighted and inherently impractical thing to do. Your position is tantamount to that of a position that Hewlett-Packard once took vis-a-vis assembly language/OS hacking on their 3000 computers: "The user can only get himself into trouble. Don't allow it. There can't possibly be any cases where the users work would be impaired by this restriction." Needless to say, a good many of us did suffer and HP did, eventually, back down somewhat on that issue. I can certainly see the benefits of preventing foolish people from bouncing copies of /etc/termcap back and forth across the atlantic, but not if it means that *correct* messages get bounced back to the user! No fix that actually breaks things that aren't broken can be considered a fix in any sense of the word. It's a BUG. Plain and simple. This is even more objectionable in light of the fact that unido is trying to become a major backbone for Germany and will soon be in the position of being able to bounce hundreds of messages a day, not just a few. My god! Progress! Now I know that your counter-argument to all of this is quite simple: "Get the map entries fixed." However, it's too simple by far. This is not BITNET or, especially, the INTERNET. Messages (or routes) do not get flashed back and forth between nodes at 56KBaud (or 9600 at the least) and it is *NOT* readily detectable when a node drops off into a black hole. "Getting the map entries fixed" can often be a highly tedious process that is only magnified by the fact that we happen to be several thousand miles away! Even when I lived in the U.S. this was not always very easy. System administrators get sick, take long vacations or are often just too overworked (or lazy) to get around to doing it. Take the even more common case of a system going off the air for a few weeks, I.E. ucbvax is hit by a bolt of lightning/decides to write over its spool directory/develops a pathological hardware bug and I happen to find this out and want to route around it. I CAN'T!! Even though ten different alternate routes lie there tantilizingly, I CAN'T USE THEM BECAUSE YOUR MACHINE WOULD RATHER ROUTE MY MESSAGE STRAIGHT TO HELL! In short, re-routing is a NICE IDEA but it's on the wrong net, with the wrong implementation strategy and done for all the WRONG REASONS by the WRONG PEOPLE. If you can offer me a way of overriding this mechanism so that I can get around dead machines, dead links and DEAD system admin- istrators then I shall raise no further fuss and quietly send my mail in peace. If you cannot, then I still must respectfully maintain that you are doing a Very Bad Thing and you should stop. I'm just trying to get my mail to go where I tell it to, is that so much to ask? I'm trying to get a direct uunet link here, but these things take time... Jordan Hubbard -- Paul Vixie Work: vixie@decwrl.dec.com decwrl!vixie +1 415 853 6600 Play: paul@vixie.sf.ca.us vixie!paul +1 415 864 7013
lear@NET.BIO.NET (Eliot Lear) (10/27/88)
Jordan and Paul: This is just a quick note to say that I saw your message and that I'll reply to it when I actually have some time. Regards, Eliot -- Eliot Lear [lear@net.bio.net]