Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watnot!watmath!clyde!rutgers!brl-adm!seismo!lll-lcc!styx!ames!oliveb!jerry
From: jerry@oliveb.UUCP
Newsgroups: news.admin
Subject: Re: Sending Email to People who post to the USENET
Message-ID: <596@oliveb.UUCP>
Date: Tue, 3-Mar-87 21:54:10 EST
Article-I.D.: oliveb.596
Posted: Tue Mar  3 21:54:10 1987
Date-Received: Fri, 6-Mar-87 04:42:45 EST
References: <1354@hplabsc.UUCP> <228@devon.UUCP>
Reply-To: jerry@oliveb.UUCP (Jerry F Aguirre)
Organization: Olivetti ATC; Cupertino, Ca
Lines: 35

In article <228@devon.UUCP> paul@devon.UUCP (Paul Sutcliffe Jr.) writes:
>Path: vu-vlsi!cbmvax!rutgers!seismo!lll-lcc!pyramid!hplabs!hplabsc!taylor
>
>While this is the path that the article took to get here, it ignores
...
>Your mail-path-parsing scheme, not knowing this, would attempt to send
>mail via:
>
>	...!seismo!rutgers!cbmvax!vu-vlsi!...

NO, you have it backwards.  Unlike reversing a UUCP address the "Path:"
line is ALREADY in the correct format for a UUCP address.  So for the
example given it would be:

	...!hplabs!hplabsc!taylor

which doesn't seem especially long to me.

I don't think the proposed algorithm should be the default way to route
mail but if the automatic routing was unable to find the target site
then it could privide an alternative.

As for the other objection, I believe for every "news but not mail" link
that exists there are 10 published connections that don't really work. 
With connections constantly changing routing is not exactly an exact
science.

Suppose you were trying to drive cross country and road maps were only
published every year or so.  Meanwhile the road crews were closing
highways for repair and finishing new roads all the time.  Wait, what am
I saying?

					Jerry Aguirre
					Systems Administration
					Olivetti ATC