Date: Tue, 19 Oct 1993 11:33:43 -0700
From: jak@violet.berkeley.edu (John A. Kunze)
Message-Id: <199310191833.LAA21912@violet.berkeley.edu>
To: drobison@library.Berkeley.EDU, hoymand@joe.uwex.edu
Subject: Re: URNs and Meta-Information - The Value of ISBN
> Date: Mon, 18 Oct 1993 10:48:47 -0700 (PDT)
> From: David Robison <drobison@library.Berkeley.EDU>
>
> As a Librarian and Internet Publisher, I see two potenetial problems
> with the current URI plan (as far as I have been able to follow the
> discussion). First, allowing the publisher to assign a single URI
> to an object in its many formats obfuscates the fact that different
> formats af a document (as well as different versions) have different
> meaning.
They have different meaning depending on the point of view of the body
assigning the URI. Libraries have well-developed cataloging rules.
Other Internet bodies will have very different rules. Right or wrong,
they may well see different formats as having the same meaning,
particularly if their customers sometimes (or always) value that point
of view (and probably other points of view as well).
> Second, allowing multiple naming authorities (ID Authorities)
> the ability to name identical objects with different URIs seems to
> contradict the purpose of the URI in the first place.
It would be nice if there were simply one ID Authority, or several ID
Authorities for strictly non-overlapping chunks of the object universe,
but I don't think that's possible.
This may be a naive example: two libraries hold a copy of "Huck Finn",
same publisher, same edition. Library one assigns an internal acquistion
number N1, library two assigns it N2. The books also have an ISBN, a
Library of Congress number and a Dewey Decimal number.
It may be argued that only the LoC number, or the ISBN, should be used
as the URN. But each of the five numbering schemes has *value*. To some
users the internal numbering scheme of their departmental secretary will
consistently have the highest perceived value. URNs need to work there.
-John