Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!cs.utexas.edu!uunet!mnetor!tmsoft!becker!bdb
From: bdb@becker.UUCP (Bruce Becker)
Newsgroups: news.newusers.questions
Subject: Re: Converting Internet addresses to UUCP addresses
Message-ID: <2233@becker.UUCP>
Date: 15 Jan 90 03:36:51 GMT
References: <325@fltk.UUCP> <2193@stealth.acf.nyu.edu> <1160@scorn.sco.COM> <1644@dsac.dla.mil>
Reply-To: bdb@becker.UUCP (Bruce Becker)
Distribution: usa
Organization: G. T. S., Toronto, Ontario
Lines: 69

In article <1644@dsac.dla.mil> nor1675@dsac.dla.mil (Michael Figg) writes:
|In article <1160@scorn.sco.COM>, davidbe@sco.COM (The Cat in the Hat) writes:
|> 
|> -that understands the UUCP map. (Try noriega%jail.usa.gov@uunet.uu.net
|> -with most mailers.) Mail routing can be very complex.
|> 
|
|What does the '%' do in an address like this? Is it saying noriega at 
|jail.usu.gov, which is a system that uunet.uu.net knows about?

	The rules say that there can be only
	one "@" in an address, since it is supposed
	to be a name and not a route. The "%" was
	invented as a kluge to add some routing
	like "!" in UUCP, and is mostly equivalent
	to "@" in syntax. 

	The problem is that once "%" gets into
	the UUCP domain, many mailers don't
	recognize it as a special character -
	'smail' is a typical example. An address
	"you%there@becker.UUCP" is presumably intended
	to go to UUCP site "becker" who will then
	figure out "you%there" and send to user
	"you" at site "there", as if "you@there"
	had been specified from there.

	Until quite recently, when I put in a patch
	to 'smail', such mail would be delivered to
	local user "you%there" on "becker" - 'smail',
	finding no such user, would bounce the message.

	The other problem is that there is no established
	precedence in operators between "@" and "!".
	"uunet!bdb@becker" is subject to interpretation
	as "becker!uunet!bdb" or "uunet!becker!bdb",
	when moved into the UUCP domain. So in order
	to address this site you might need to try
	"becker!bdb@uunet" to get there instead,
	depending on local or intermediate mailer
	implementations. For 'smail', the second form
	turns out to be the correct one.

	When "%" is interpreted as "@", the same problems
	arise. They are slightly simpler due to the fact
	that "%" is not official, so the default behavior
	of 'smail' above is technically correct, if
	unforgiving. If "%" is interpreted, it is
	done last, after all other mistakes are made
	(if any), so it does have an implied precedence
	of a sort.

	This is lotsa fun, but not necessarily productive
	when you want to get mail from one network to
	another - "%" can help, but like "!", you end
	up needing to know local and intermediate conditions
	in order to use it effectively.

	I've probably said things which others will have
	differing opinions on - if so, the ensuing
	brouhaha will likely be educational, if things
	are working as they should...

Cheers,
-- 
  ,,,,	 Bruce Becker	Toronto, Ont.
w \$$/	 Internet: bdb@becker.UUCP, bruce@gpu.utcs.toronto.edu
 `/c/-e	 BitNet:   BECKER@HUMBER.BITNET
_/  >_	 "Money is the root of all money" - Adam