The Semantic Web - Social Semantics: The Search for Meaning on the Web - page 64

3.3 Information and Non-information Resources

One question is whether or not there should be some way to distinguish between
URIs used to access web-pages and Semantic Web documents, and URIs used as
names for things like physical entities and abstract concepts that are not 'on the
Web.' This latter class of URIs, URIs that are used as names for entities and abstract
concepts, are called Semantic Web URIs . Should a URI be able to both name a non-
Web accessible thing in addition to accessing a representation of the thing? This is
a difficult question, as it seems the class of web-pages and physical people should
be disjointed (Connolly 2006). The W3C TAG took on this question, calling it the
httpRange-14 issue, which was phrased as the question: what is the range of the
HTTP dereference function? (Connolly 2006).
The TAG defined a class of resources on the Web called an 'information
resource,' which is a resource “whose essential characteristics can be conveyed in a
message” (Jacobs and Walsh 2004). In particular, this means that an information
resource is a resource that can be realized as an information-bearing message,
even with multiple encodings. A resource is defined by its sense (content), not
the encoding of its Web representations. So information resources would naturally
include web-pages and so resources on the hypertext Web, as well as most digital
things. However, there are things that cannot be realized digitally by a message ,
but only described or depicted by digital information. These things are non-
information resources . Their only realization is themselves. Many analogue things
therefore are non-information resources. It appears that this distinction between
information resources and non-information resources is trying to get at the heart
of the distinction between a resource being a web-page about the Eiffel Tower and a
resource for the Eiffel Tower itself. A web-page is an information resource, but the
Eiffel Tower itself is a non-information resource, as is the text of Moby Dick or the
concept of red.
The distinction is more subtle than it first appears. The question is not whether
something is accessible on the Web, but whether it can be accessible on the Web
by being in theory transmitted as an encoding, and therefore Web representation,
in a message. For example, imagine a possible world where the Eiffel Tower does
not have a web-page. In this world, it would seem counter-intuitive to claim that
the web-page of the Eiffel Tower is then not an information resource just because
it happens not to exist at this moment. This is not as implausible as it sounds, for
imagine if the Eiffel Tower's web server went down, so that http://www.tour-eiffel.fr
returned a 404 status code. A more intuitive case is that of the text of Moby Dick .
Is the text of Moby Dick an information resource? If the complete text of Moby Dick
isn't on the Web, one day it might be. However, a particular collector's edition of
Moby Dick could not be an information resource, since the part of that resource isn't
the text, but the physical topic itself. Are ordinary web developers expected to have
remarkably scholastic discussions about whether or not something is essentially
information before creating a Semantic Web URI?
Both a web-page about the Eiffel Tower and the text of Moby Dick are, on some
level of abstraction, carrying information about some content in some encoding.