Message-Id: <9305161823.AA00336@mocha.bunyip.com>
To: Chris Weider <clw@merit.edu>
Subject: Re: Internet Draft on URNs
In-Reply-To: Your message of Fri, 14 May 1993 18:03:11 -0400.
<9305142203.AA23447@merit.edu>
Date: Sun, 16 May 1993 14:23:09 -0400
From: John Curran <jcurran@nic.near.net>
--------
]
] 3.2 What URNs are *not*
]
] URNs are not intended to be human-readable in the sense that a human could
] look at the URN and determine anything about the contents of the resource.
] While the Naming Authority (q.v.) has the final determination of the contents
] (subject to the syntax constraints), the Naming Authority is STRONGLY
] discouraged from placing metainformation about the resource into the resource's
] URN, as the URNs are not expected to be read, and because this paper will
] specify only four consistent components of the URN. Although there have been a
] number of proposals placing extensive semantics on the contents of the URN
] [Spero 1992, Kunze 1993], it was decided by the authors of all the proposals
] that all metainformation should be conveyed using another mechanism, and that
] the Naming Authority should assume that humans will never look at the contents
] of the URN to determine qualities of the resource they are retrieving, and
] would not be required to guess from a given URN the URN of a document which
] might be related.
I agree with the above, but note that _each_ naming authority may name its
documents as it sees fit within the constraints of syntax. If people find
the URNs of certain naming authorities functionally challenged, then I am
confident that sufficient backpressure will exist on the naming authority
to reevaluate its chosen naming methodology.
Minor nit: "Naming Authority should assume that humans will never look at the
contents of the URN to determine qualities of the resource..."
We may want to add a note to the effect that Naming authorities should
expect that _some_ folks will attempt to examine at the contents of the
URN to attempt to determine origin of the resource, and that some
classes of URNs may support this (e.g. URN:ietf.isoc.org::wg-uri-urn-00).
While this is not a requirement of a URN, naming authorities should be
aware of the preexisting expectation.
/John