ambar@bloom-beacon.mit.edu (Jean Marie Diaz) (07/31/89)
From: les@chinet.chi.il.us (Leslie Mikesell)
Date: 17 Jul 89 19:12:36 GMT
Did anyone think that people typed those 20-hop paths by hand? Especially
people who don't know what they are doing? Replies to news articles
and subsequent replies to those messages are almost certainly the source
of everything the rabid rerouters complain about.
So, why not fix it in the news software by making hosts that are capable
of re-routing rewrite the Path: line so that mail following the Path:
will be "dynamically" re-routed? That way the dumber downstream sites
don't need to change anything (any they probably won't anyway).
Because one of the purposes of the Path: header in news (in my opinion,
the PRIMARY purpose) is to track the places where that article has
already gone, and not send it on further. Your suggestion would have
news travelling in loops, further overloading high-bandwidth sites.
Using the Path: header as a mail reply path is, at best, unreliable, and
at worst (think of moderated newsgroups, or newsgroups that are
gatewayed to and from mailing lists) completely wrong.
AMBAR
ambar@bloom-beacon.mit.edu {mit-eddie,uunet}!bloom-beacon!ambarjbuck@epimass.EPI.COM (Joe Buck) (07/31/89)
> From: les@chinet.chi.il.us (Leslie Mikesell) > Date: 17 Jul 89 19:12:36 GMT > > Did anyone think that people typed those 20-hop paths by hand? Especially > people who don't know what they are doing? Replies to news articles > and subsequent replies to those messages are almost certainly the source > of everything the rabid rerouters complain about. Yes, this is probably true. But the problem is that the Path: header was never designed to be the return address; the Reply-To: header should be used, and if that header isn't present, the From: header should be used. When building and installing news, INTERNET should always be defined. Even on dumb systems, the "mailpaths" file may be used to cause replies to articles to be routed to a smart host for routing. > So, why not fix it in the news software by making hosts that are capable > of re-routing rewrite the Path: line so that mail following the Path: > will be "dynamically" re-routed? That way the dumber downstream sites > don't need to change anything (any they probably won't anyway). SUFFER, "dumber downstream sites"!! :-) Or clean up your act. In article <AMBAR.89Jul30161342@binkley.mit.edu> ambar@bloom-beacon.mit.edu (Jean Marie Diaz) writes: >Because one of the purposes of the Path: header in news (in my opinion, >the PRIMARY purpose) is to track the places where that article has >already gone, and not send it on further. Your suggestion would have >news travelling in loops, further overloading high-bandwidth sites. This isn't "one of the purposes"; it's THE purpose. Actually, without it, articles would not travel in loops, because articles have message-IDs (and duplicates would be rejected). But dropping the Path: mechanism would greatly increase the number of duplicate articles arriving at what we used to call backbone sites. This wouldn't cause The Death Of Usenet As We Know It (TDOUAWKI) but it would greatly increase some folks' phone bills. >Using the Path: header as a mail reply path is, at best, unreliable, and >at worst (think of moderated newsgroups, or newsgroups that are >gatewayed to and from mailing lists) completely wrong. Right on. Add: totally unnecessary; any site can install a free routing mailer, or pass news replies to a neighbor who has one. Anyone who sends mail on a 20-hop path deserves to wait a long time for it to be delivered. -- -- Joe Buck jbuck@epimass.epi.com, uunet!epimass.epi.com!jbuck
clewis@eci386.uucp (Chris Lewis) (08/01/89)
In article <3466@epimass.EPI.COM> jbuck@epimass.EPI.COM (Joe Buck) writes: >> Did anyone think that people typed those 20-hop paths by hand? >Yes, this is probably true. But the problem is that the Path: header >was never designed to be the return address; the Reply-To: header >should be used, and if that header isn't present, the From: header >should be used. When building and installing news, INTERNET should >always be defined. Even on dumb systems, the "mailpaths" file may be >used to cause replies to articles to be routed to a smart host for routing. If you have a smart mailer to begin with, you can also use rn's INTERNET configuration option. Or, even hand-cruft a mod to Rnmail that does path lookups and replaces the "To:" line (it's a pretty minor mod). The path lookups can be as simple as finding out whether the machine is mentioned in uuname's output, and if not, replacing the whole shebang with <pathtosmartmailer>!<destinationsite>!user. The facilities exist to avoid the 20 hop syndrome: SA's should make sure that they read the associated documentation when installing news. -- Chris Lewis, R.H. Lathwell & Associates: Elegant Communications Inc. UUCP: {uunet!mnetor, utcsri!utzoo}!lsuc!eci386!clewis Phone: (416)-595-5425
ckd@bu-pub.bu.edu (Christopher K Davis) (08/02/89)
In article <*@eci386.uucp> clewis@eci386.uucp (Chris Lewis) writes: In article <3466@epimass.EPI.COM> jbuck@epimass.EPI.COM (Joe Buck) writes: >> Did anyone think that people typed those 20-hop paths by hand? >Yes, this is probably true. But the problem is that the Path: header >was never designed to be the return address; the Reply-To: header >should be used, and if that header isn't present, the From: header >should be used. When building and installing news, INTERNET should >always be defined. Even on dumb systems, the "mailpaths" file may be >used to cause replies to articles to be routed to a smart host for routing. [chomp] The facilities exist to avoid the 20 hop syndrome: SA's should make sure that they read the associated documentation when installing news. The *really* fun stuff happens with NNTP. Look at my path header--unless some programming elf changed something, it will end with "...!bu-cs!ckd". I'm not posting from bu-cs, I'm posting *via* bu-cs... and I don't even have an account there. Just wait until someone tries to R to this message, and their sysadmin left it using paths for return mail... Is it any wonder why my .sig tells people to ignore the header? Chris Lewis, R.H. Lathwell & Associates: Elegant Communications Inc. UUCP: {uunet!mnetor, utcsri!utzoo}!lsuc!eci386!clewis Phone: (416)-595-5425 -- /\ | / |\ @bu-pub.bu.edu <preferred> | Christopher K. Davis, BU SMG '90 / |/ | \ %bu-pub.bu.edu@bu-it.bu.edu | uses standardDisclaimer; \ |\ | / <for stupid sendmails> | BITNET: smghy6c@buacca \/ | \ |/ @bucsb.UUCP <last resort> or ...!bu-cs!bucsb!ckd if you gotta. --"Ignore the man behind the curtain and the address in the header." --ckd--