Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!yale!think!barmar
From: barmar@think.com (Barry Margolin)
Newsgroups: comp.mail.misc
Subject: Re: configurable delimiters needed in aliases file for x.400 addresses
Message-ID: <40966@think.Think.COM>
Date: 24 Jul 90 03:40:08 GMT
References: <1990Jul22.123110.1198@tolerant.com> <40897@think.Think.COM> <15691@bfmny0.BFM.COM>
Sender: news@Think.COM
Organization: Thinking Machines Corporation, Cambridge MA, USA
Lines: 61

In article <15691@bfmny0.BFM.COM> tneff@bfmny0.BFM.COM (Tom Neff) writes:
>What the Net needs is Directory Assistance -- an automated entity to
>ask for machine addresses.

OK, supposing we implement an online Directory Assistance, we'd have to
define a query protocol (in your example, email was the transport
mechanism).  There would have to be a syntax for writing the queries.  It
could be an encoding of something like SQL:

	SELECT ADDRESS
	WHERE Company_Name="Thinking Machines"
	& Department="Marketing"
	& Role="Vice President"

Sure looks like what X.400 addresses encode.

And that's essentially what X.500 is -- an automated directory assistance.
It takes the complicated X.400 addresses (they're actually called ORNames,
short for Originator/Recipient Names), looks up the attributes in a
distributed database, and figures out the mail path.

Rather than sending a query to the directory service and waiting for the
response, you simply put all the information that would be in the query
right into the envelope, and the directory service is invoked automatically
and the message is sent to its destination.  If the information is
ambiguous I'd expect to receive a failure notice describing the ambiguity
(I don't know whether this is actually in the X.500 protocol -- I am not
actually familiar with the implementation details of any of the stuff I've
been talking about in this chain).  But in cases where I supply enough
information, why should I have to wait for a response so I can send the
real message?  I might be about to leave for the evening, and I'd like to
send a message so that it will hopefully reach its recipient by the time he
reads his mail in the morning, so I don't want to have to sit around,
perhaps for hours if mail is backed up, waiting for the information (the
reason telephone D.A. is acceptable is because it rarely takes more than a
couple of minutes to use).

Now, it would be nice if, when I sent a message that uses a long ORName,
the mail system would automatically send back a message that says "in the
future, you can use this abbreviated ORName."

Someone else suggested sending mail to Postmaster@domain, asking the
postmaster to forward the message.  First of all, most email postmasters
are system administrators and/or system programmers, not receptionists.
Second, top executives might not appreciate their mail being sent to
underlings (sure, many postmasters have access to the mail queues, but they
can usually be trusted not to use this to violate confidentiality, whereas
it's easy to accidentally read someone else's mail if it's sent to you).
Even if the mail simply asks the postmaster to send back someone's email
address, it's not really the postmaster's job.  I'm not even our
postmaster, but I've received mail from random people saying "I saw from
your posting that you're at Thinking Machines, and I know So-and-so works
there, could you please send me his email address?"  It happens rarely
enough that it's not annoying, but if it happened every day I'd quickly
tire of it.  In the very least, there should be a different standard
address for this (Addressmaster@domain).
--
Barry Margolin, Thinking Machines Corp.

barmar@think.com
{uunet,harvard}!think!barmar