Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site umcp-cs.UUCP
Path: utzoo!linus!philabs!cmcl2!seismo!lll-crg!gymble!umcp-cs!chris
From: chris@umcp-cs.UUCP (Chris Torek)
Newsgroups: net.mail
Subject: Re: How to put UUCP hosts into world domain tree?
Message-ID: <1616@umcp-cs.UUCP>
Date: Tue, 17-Sep-85 17:21:23 EDT
Article-I.D.: umcp-cs.1616
Posted: Tue Sep 17 17:21:23 1985
Date-Received: Fri, 20-Sep-85 01:52:55 EDT
References: <1469@cbosgd.UUCP> <12532@styx.UUCP>
Distribution: net
Organization: U of Maryland, Computer Science Dept., College Park, MD
Lines: 71

[Warning:  Human (and Elvish of course :-) ), not software, issues
discussed herein.  If you are not interested in such, do not read on.]

In article <12532@styx.UUCP> mcb@styx.UUCP (Michael C. Berch) writes:
> In article <1469@cbosgd.UUCP> mark@cbosgd.UUCP (Mark Horton) writes:
>> [possibilities include trying to] fit into the ARPA organizational
>> space, with names like cbosgd.ATT.COM and ucbvax.Berkeley.EDU.

> This alternative looks like a clear winner to me. ...  The idea of
> uniting the domain and name spaces of North America's two most
> visible electronic mail internets (DDN and UUCP) is tremendously
> attractive, even though it might mean giving up some measure of
> domain identity to the people who will eventually administer the
> entire domain name scheme for DOD.

That is indeed the problem, as Mark points out:

>> This might not be hard, given the facilities of smail and pathalias,
>> but we do not have official permission from ARPA to do this, and
>> they have not been moving quickly to resolve the problem.  (Since
>> they keep the registries for EDU, COM, GOV, MIL, and ORG, we would
>> have to register with them.)

It would not actually be necessary to register with them; however,
this would likely create an administrative disaster.  The key
feature of domains, as far as DARPA is concerned, is that administrivia
is moved out of one central region, yet is still clearly defined:
it is somewhere written that UMD.EDU shall be the responsibility
of the University of Maryland.

While we could simply ``take over'' the name space so thoughtfully
provided us, it would create a great deal of confusion among many
sites.  No longer would ``*.COM'' be found in the COM registry on
the ARPAnet, for ATT.COM would exist only in the UUCP world.  I do
not believe that the NIC would be happy about this situation; ARPA
sites that receive UUCP mail and are unable to reply (as they are
not running the UUCP routing software so cannot understand why the
COM registry does not know about ATT) would likely bombard them
with queries.

This, then, is the problem:  How can we adopt an existing namespace
without hurting those already using it?  The answer, I think, is
that we cannot.  The UUCP routing system is sufficiently different
from the ARPA system that it will not simply ``plug in''.  From a
user's viewpoint, however, merging the namespaces is very attractive.
If we do this, we must decide which naming scheme we shall use.
I think that, in terms of software, it actually matters not whether
we use ATT.UUCP or ATT.COM; the software will handle either, an it
becomes common usage.  It is true that a single special case is,
at least, more obvious, and for this reason we should choose
ATT.UUCP.  But a UUCP domain has no real administrative implications,
unlike other domains, so this is perhaps misleading, and we should
then choose ATT.COM.

				* * *

Elves seldom give unguarded advice, for advice is a dangerous
gift, even from the wise to the wise, and all courses may run ill.
But if you demand advice, I will for friendship's sake give it.
 . . .  Oh, sorry, that was a different conversation.

I personally believe that we should adopt the existing namespace
wholeheartedly.  d.osg.cb.att.com, ucbvax.berkeley.edu, and the
like are at least consistent with their ARPA-world counterparts.
But I have not invested as much time in studying these matters as
others have, and I think we should hear their viewpoints before we
commit ourselves to any one course of action.
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 4251)
UUCP:	seismo!umcp-cs!chris
CSNet:	chris@umcp-cs		ARPA:	chris@maryland


Brought to you by Super Global Mega Corp .com