Date: Thu, 17 Feb 94 14:03:39 PST
From: redback!jimmc@eskimo.com (Jim McBeath)
Message-Id: <9402172203.AA28955@redback.>
To: uri@bunyip.com
Subject: Unresolvable URNs
I have been trying to understand how URNs will actually be used by
comparing them to how references are generally used in printed media.
I don't know if this is a valid comparison, but it seems like there
must be something we can extract from the years of experience people
have using printed references.
Karen Sollins writes:
> Resolution: It is quite possible that there will be URNs that
> cannot be resolved for all time. Two reasonable situations in which
> this may occur if a URN is created as a place holder for a resource
> that does not yet exist or in the case that a resource is either
> temporarily or permanently inaccessible (deleted, killed, etc.) In
> this latter case, the fact that there was/is a reference may be
> important, even if the resource is not currently accessible.
In printed media, a reference to another document generally includes
not only an opaque unique identifier (such as an ISBN number), but
a small amount of additional information, such as Author and Title.
Theoretically, only the ISBN is required, from which the referenced
document can be retrieved. But what if the reference can not be
resolved? With only an opaque identifier, the reader can get no information
about the contents of the referenced document. With a name and an author,
it is at least possible to have an inkling of what the document is about,
and to attempt to chase down other related references, for example by
the same author.
Here's an example scenario: a naming authority has determined that ascii and
postscript versions of a document are not identical, and has thus assigned
them different URNs. The URN for the postscript version gets put into a
document, but then the postscript version gets destroyed, so is no longer
available. When reading the document containing the URN pointing to the
postscript version, I am no longer able to access the contents of that
postscript document. If I had just a little bit of information about the
referenced document, I might be able to do some searching and find the ascii
version of it.
You could argue that the naming authority is responsible for remembering the
relationship between the ascii and postscript versions, and thus should be
able to find the one when given a reference to the other. I suspect
this would not work, both because people would forget to tell this to
the system, and because URNs will almost certainly be resolved by authorities
other than the original naming authority (such as a local or regional
caching site), and those other sites will probably not have the meta-info
relating documents that the original naming authority has. It's possible
that the original naming authority will become defunct, but other sites
which already have documents could still resolve URNs from that naming
authority which reside at that site.
I realize that this information is not strictly part of the URN, and is
more like the URC stuff that was discussed a while back. What I'm curious
about is whether people believe there will be any kind of "accepted usage"
for URNs so that URC-type information is somehow included in the same
document that includes the URNs, just in case the document referenced by
the URN is, for whatever reason, unavailable. This may not be important
for short-lived documents, but may be more important for a document
expected to live for a hundred years, during which time it seems
reasonably likely that some naming authorities may become defunct.
-Jim McBeath
jimmc@eskimo.com