home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!gatech!concert!duke!wolves!news
- From: news@wolves.Durham.NC.US (The Wolfe of the Den)
- Newsgroups: comp.mail.uucp
- Subject: Re: More problems with the UUCP maps
- Message-ID: <C19HML.58z@wolves.Durham.NC.US>
- Date: 22 Jan 93 15:27:56 GMT
- References: <C10os2.E8B@cs.psu.edu> <1993Jan19.124933.1397@globv1.hacktic.nl> <1jkiuaINNlb4@usenet.INS.CWRU.Edu>
- Organization: Wolves Den UNIX
- Lines: 60
-
- trier@slc6.ins.cwru.edu (Stephen C. Trier) writes:
- >peter@globv1.hacktic.nl (Peter Busser) writes:
- >>The underlying assumption you made is that every mail package returns bounces
- >>when it can't find the addressed user. Are there any statistics that support
- >>this assumption?
- >
- >Experience, I'd say. I'd consider a UUCP package that doesn't return
- >bounces to be broken.
- >
- >What possible reason could one have for _not_ bouncing messages to non-
- >existent users?
- >
- >The only package I know of that doesn't issue bounces is UUPC. Its ports
- >to other systems may share this deficiency. The reason is either that it
- >is a "single user package" or that it is not worth the effort to implement.
-
- Well, since UUCP *per se* is NOT a mailer, but a simple file
- transfer protocol for point-to-point dial up connections, UUCP has no
- business trying to be a mailer *per se*
-
- Now, most UUCP packages *should* have the capability to respond
- to the sender when a uucp request gets hung in the queue and doesn't get
- dispatched in a reasonable length of time. In this case, the UUCP
- cleanup daemon looks to see if it is a mail message, and tries to inform
- the original sender as well as the local administration.
-
- >I'm not aware of any Unix UUCP mailers that will not issue bounces, but
- >all the folks around here seem to run sendmail, smail2.5 or smail3.x, all
- >of which generate correct bounces.
- >
- >A more likely explanation is that there are enough path-manglers and flawed
- >rabid rerouters out there that the messages got lost on the way there and
- >the bounces on the way back.
-
- You forget also that a UUCP link is not necessarily
- bi-directional! Generally, if a site polls another, messages going in
- both directions can be exchanged, but there are ways to break the
- symmetry, and there may not be a mailer on the reverse link that
- understands uucp syntax.
-
- Additionally, there are some sites that allow anonymous UUCP
- connects with "rmail" privledge, which would allow any site to dump mail
- bound for some site into the stream, but there would be no reverse
- channel.
-
- This can be quite frustrating too. Someone may dump mail into
- your site via the anon uucp connect, and then wonders why you don't
- answer their email request. :-(
-
- As another example, the recent flak over the link between
- zuma.UUCP and antatolia was set up so that zuma could dump news into the
- net, but not have a useable reverse channel. This "convenient"
- arrangement allows Serdar Argic to propagandize the net while not being
- subject to any form of peer or social pressure for doing so. (And this
- arrangement also violates the existing Usenet/NetNews RFC's.)
- --
- Usenet Net News Administrator @ The Wolves Den (G. Wolfe Woodbury)
- news@wolves.durham.nc.us news%wolves@cs.duke.edu ...duke!wolves!news
- "The flame war is a specific Usenet art form." --me
- [This site is not affiliated with Duke University. (Idiots!) ]
-