Revision of Chris & Peter's URN paper

[Chris Weider <[e-mail]>]

Gang:
  Here is a revision of the URN draft that we submitted earlier this year. It
has a number of changes from the earlier draft. In addition, Peter has seen
this revision but has not yet had an opportunity to comment; Alan suggested
that I send it out anyway in time for Houston.

Changes from earlier draft:

1) The multiple colon syntax has been removed; individual fields are now
separated by colons, with positional semantics indicating the individual 
fields and a 'colon-counting' parse technique.

2) The characters < and > are use as delimiters for the URN. Unlike the
URL, this is built into the syntax of the URN; I still feel there is a need 
for termination characters especially when these are going to be cut and 
pasted.

3) (Major change) A fifth field has been added, specifying the encoding
scheme used for the opaque string. It seems to me that it would be unwise to
limit the potential character sets usable for the opaque string, particularly
if we're planning for the future. However, we still have those pesky mailers
to deal with... so I have defined one encoding scheme, ASCII encoded ASCII.
While this may seem rather redundant, think about other possibilities such
as ASCII encoded UNICODE, ASCII encoded binary checksums, etc.

Points emphasized more in the current draft:

1) The fact that the naming authority identifier may be hierarchical in 
nature, and multi-leveled. Apparently this was not brought out enough in the
previous draft as we kept getting questions about it. 

2) The fact that the primary function of the URN is to provide a persistant 
(location independent) identifier.

Controversial points still in the draft:

1) The fact that the naming authorities have complete control over the opaque
string assignment, BUT that they are encouraged not to have semantically
meaningful subparts in the opaque string, which would give the human reader
an indication of its use. I sent out a message about the dangers of placing
comparative attributes (such as version number) in the URN itself; since I
didn't get any response, I must assume that everyone agrees with me 
(wry grin)...

Well, with that, here's the latest draft.
Chris


INTERNET--DRAFT						Chris Weider
IETF URI Working Group					Merit Network, Inc.
							Peter Deutsch
							Bunyip Information
							  Systems, Inc.
							October, 1993

		Uniform Resource Names

Status of this Memo

In this paper, the authors propose an identifier, called the Uniform Resource
Name (URN), which is designed to provide persistent naming for resources
and objects on the Internet.

        This document is an Internet Draft.  Internet Drafts are working
        documents of the Internet Engineering Task Force (IETF), its Areas,
        and its Working Groups.  Note that other groups may also distribute
        working documents as Internet Drafts.

        Internet Drafts are draft documents valid for a maximum of six
        months. Internet Drafts may be updated, replaced, or obsoleted
        by other documents at any time. It is not appropriate to use 
        Internet Drafts as reference material or to cite them other than
        as a "working draft" or "work in progress."

        Please check the I-D abstract listing contained in each Internet
        Draft directory to learn the current status of this or any 
        other Internet Draft.

	This Internet Draft expires March 20, 1994.


1: Introduction

A Uniform Resource Name (URN) is an identifier which can be used to uniquely
identify a resource, and is designed to provide persistent naming for 
networked objects.  This name would stay the same no matter what the 
current location(s) of the object was.

2: Motivation

This work comes out of the discussions held at the Uniform Resource Identifier
meetings at the IETF, and from further discussions among interested parties.
Currently, the only standard identification scheme for resources on the Net is
the Uniform Resource Locator (URL) [Berners-Lee 1993]. This "Locator"
is designed to provide a uniform way of specifying location and retrieval 
information for networked objects. The URL, however, will not provide a stable,
long-lived reference to a resource as the resources have a bad habit of moving 
out from under the locator. Also, a given resource may have multiple URLs if
it resides at a number of different locations on the net, or is available
under a number of different access methods. Thus it is difficult to tell,
given two different URLs, whether the resources they point to are the same
or different without retrieving both of them. The Uniform Resource Name, or 
URN, has been designed to alleviate these problems.


INTERNET--DRAFT		Uniform Resource Names			Weider, Deutsch

3: The Uniform Resource Name (URN)

3.1 Functionality

The URN is designed to provide persistent naming for objects on the net. It
is intended to be used in conjunction with a directory service, which can 
provide a URN -> URL mapping [Weider 1993]. This URN-URL architecture allows 
permanent references to be made to resources without worrying about their 
current locations. It is also intended to provide some detection of duplicates 
in responses to queries of various resource location services. 

3.2 What URNs are *not*

URNs are not required 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 five 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.

3.3 Components of the URN

There are five components to the URN, separated by colons; the keyword
'URN', a code specifying the character set encoding of the rest of the URN, a 
naming authority scheme identifier, a naming authority identifier, and an 
opaque string. The URN is surrounded by the characters '<' and '>', which are 
part of the syntax. Each part is described below. No component of the URN 
can contain the characters ':', '>', ' ', or '\'  unless they are escaped by a 
backslash character '\'. 

3.3.1 URN examples

<URN:ASCII:IANA:merit.edu:1929642>

<URN::ISBN_Publisher_ID:0_201_12:xyzx\:mnopq>

<URN:ASCII:IANA:12456:1\:<\>\:2345>

3.3.2 The character set encoding code

This string identifies the encoding scheme used for the rest of the URN. 
There is only one defined at this time, ASCII, which indicates that the
rest of the URN is ASCII encoded ASCII. If this component is empty,
the default encoding scheme is assumed. The default encoding scheme is
ASCII.


INTERNET--DRAFT		Uniform Resource Names			Weider, Deutsch

3.3.3 The naming authority scheme identifier

The naming authority scheme identifier is a string which is the name of a 
protocol or organization which guarantees the uniqueness of the naming 
authority identifier which follows. Naming authority scheme identifiers defined 
at this time are 

	IANA
	ISBN_Publisher_ID

3.3.4 The naming authority identifier

This string, along with the naming authority scheme identifier, identifies a 
naming authority that may assign URNs to resources. This string may have 
internal syntax depending on the naming authority scheme identifier associated 
with it; for example, the naming authority identifier space associated with IANA 
may be hierarchical and multi-leveled.

3.3.5 The Opaque String

The opaque string component of the URN is any string the Naming Authority 
wishes to assign to a given resource, subject only to the constraints of the 
character encoding scheme. 

As mentioned above, the Naming Authority should not assume that a 
human will ever read the URN. Also, the Naming Authority, in assigning an
opaque string to a given resource, should keep the following guidelines in 
mind:

  	1: A given opaque string should be case-insensitive (for compatibility
	   with very old systems).

	2: A given opaque string, once assigned, should never be reused. These
	   are expected to be persistent names for resources (think in terms
	   of decades).

	3: In assigning an opaque string, and thus creating a URN, the Naming
	   Authority should make provisions for a URN -> URL mapping
	   function. This need be nothing more than finding an organization
	   which is already providing this service for other URNs and making
	   arrangements to have them translate for the new URN, or could
	   be as involved as creating a new software agent to provide this
	   service.  Remember that a name is no good without some way of 
	   getting a location.

	4: URNs will be returned as pointers from a resource location service.
	   (See [Weider 1993]). Consequently, a Naming Authority should give
	   some thought to the assignation of new URNs for resources which
	   are derived in some fashion from other resources to which that
	   Authority has already assigned URNs. For example, should the
	   Postscript version and the ASCII version of a paper have the
	   same URN? While there are no universally applicable answers to
	   questions like these (for example, should the Russian and English
	   versions of a scientific paper have the same URN?) an Authority
	   should keep in mind that users will want to weed out duplicate
	   resources in the lists of URNs returned by a resource location
	   service, and consequently will be doing a lot of equality testing
	   on the URNs.

INTERNET--DRAFT		Uniform Resource Names			Weider, Deutsch

	
4: Setting up as a Naming Authority

There are 2 scheme identifiers listed here; others will no doubt be suggested
and added as this draft circulates. They are:

		IANA
		ISBN_Publisher_ID

To set one's organization up as a Naming Authority, one can use the ISBN 
publisher ID one has been assigned, or one can apply for an Enterprise
Number from the IANA (Internet Assigned Number Authority) if the organization
does not already have one. The general syntax is listed in section 5.

5: Syntax

Below is a BNF like description of the syntax of the URN. Spaces have 
been used here to separate components for readability, spaces are NOT ALLOWED
in a syntactically correct URN unless they are escaped with the '\' character. 
Square brackets '[' and ']' are used to indicate optional parts; 
a vertical line "|" indicates alternatives.  Single letters and digits stand 
for themselves. All words of more than one letter are either expanded further 
in the syntax or represent themselves.


urn              <URN: Encoding_Scheme:Authority_Id : opaque_string >

Authority_Id		Scheme_ID  : [Individual ]
Scheme_ID		IANA | ISBN_Publisher_ID | ISSN
Individual		xalphas
xalphas			xalpha [ xalphas ]
xalpha			a | b | c | d | e | f | g | h | i | j | k | l |
			m | n | o | p | q | r | s | t | u | v | w | x |
			y | z | A | B | C | D | E | F | G | H | I | J |
			K | L | M | N | O | P | Q | R | S | T | U | V |
			W | X | Y | Z | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
			9 | 0 | - | _ | . | @

The allowed characters in the opaque string are determined by the character 
set encoding code. For the code ASCII, the allowable characters are xalphas as 
above, the character ':' encoded as '\:', the character '>' encoded as '\>', and 
the character '\' encoded as '\\'.

6: References

[Kunze 1993]  Kunze, John, Resource Citations for Electronic Discovery and 
	      Retrieval, March, 1993. Circulated to ietf-uri mailing list.

[Spero 1992]  Spero, Simon, Uniform Resource Numbers, November 1992. 
	      Circulated to ietf-uri mailing list.

[Weider 1993] Weider, Chris and Deutsch, Peter. A Vision of an Integrated 
	      Internet Information Service, March, 1993. Available as

ftp://nic.merit.edu/documents/internet-drafts/draft-ietf-iiir-vision-00.txt



INTERNET--DRAFT		Uniform Resource Names			Weider, Deutsch


7: Author's addresses

Chris Weider
[e-mail]
Merit Network, Inc.
2901 Hubbard, Pod G
Ann Arbor, MI 48109
Phone: [phone]
Fax: [phone]

Peter Deutsch
[e-mail]
Bunyip Information Systems
310 St-Catherine St West
suite 202,
Montreal, Quebec H2X 2A1
CANADA