Re: UR Accessors

Daniel LaLiberte (liberte@ncsa.uiuc.edu)
Fri, 16 Jun 95 10:26:06 CDT


Date: Fri, 16 Jun 95 10:26:06 CDT
From: liberte@ncsa.uiuc.edu (Daniel LaLiberte)
Message-Id: <9506161526.AA28509@void.ncsa.uiuc.edu>
To: uri@bunyip.com
Subject: Re: UR Accessors

	At 7:06 AM 6/14/95, Daniel LaLiberte wrote:
	>We need a new name for a URI concept that has been bouncing around
	>under the name "URC0".  A URC0 has been thought of as a set of URLs
	>(previously called a list of URLs) where any one of the URLs would do
	>as the identifier of the resource.  I want to generalize this further,
	>and distinguish it from the other URC issues.

	From: ietf-lists@proper.com (Paul Hoffman)
	Please note that the URC0 draft (draft-ietf-uri-urc-trivial-00)
	describes a URC0 response as a list of URLs, *each of which can have
	metainformation*.  That's much more than "a set of URLs".

Yes, that's fine.  Later in my posting I suggest that metadata might
be associated with each URL, but that is still different from the metadata
that is associated with the abstract object as a whole.

	Further, the URLs are not equivalent: they would not all "do as the
	identifier of the resource." A URC can be for a multi-volume work. For
	example, one URC0-style URC for an encyclopedia could contain one URL
	for each volume. In this case, the metainformation for each URL would
	become quite important, no?

The metadata would be quite important, yes.  But this kind of URC0 is
quite different from the idea of an accessor that I was trying to get at.
An object that represents various ways to access the same object is 
distinguishable semantically from an object that represents how to access
the components of a composite object, even if they are structurally similar.

	I'm not sure why it is "not as much as general metadata". 

Well, I am strongly sympathetic to the idea that there is a continuum
between this accessor kind of metadata and all the rest.  Furthermore,
your example of a URC0 for an encyclopedia blurs the distinction between
data (a composite object) and metadata.  Other examples that fall in the
middle are the URLs for alternative reprentations of the same object,
and the URLs for the versions of an object.

Larry's criticism that we should not name something before we understand
it is debatable, over beer.  But the same concept of an accessor is called
operational or procedural metadata elsewhere.

I'm in favor of splitting off further metadata discussion to a different
working group and mailing list.  Stu Weibel is organizing a BOF for the
next IETF on metadata.

Daniel LaLiberte (liberte@ncsa.uiuc.edu)
National Center for Supercomputing Applications
http://union.ncsa.uiuc.edu/~liberte/