Architecture of theWorld WideWeb - Social Semantics: The Search for Meaning on the Web - page 28

should not attempt to infer properties of the referenced resource” (Jacobs and Walsh
2004). So, just because a URI says http://www.eiffel-tower.com does not mean it
will not lead one to a web-page trying to sell one cheap trinkets and snake oil, as
most users of the Web know. Second, a URI has an owner. The owner is the agent
that is accountable for a URI . Interestingly enough, the domain name system that
assigns control of domain names in URIs is a legally-binding techno-social system,
and thus to some extent a complex notion of accountability for the name is built into
URIs. Usually for URI schemes such as HTTP, where the hierarchical component
begins with an authority, the owner of the URI is simply whoever controls that
authority. In HTTP, since URIs can delegate their relative components to other users,
the owner can also be considered the agent that has the ability to create and alter
the information accessible from the URI, not just the owner of the authority. Each
scheme should in theory specify what ownership of a URI means in context of the
particular scheme.

2.2.4 Resources

While we have explained how a URI is formed, we have yet to define what a
URI is. To inspect the acronym itself, a Uniform Resource Identifier (URI) is
an identifier for a 'resource.' Yet this does not solve any terminological woes,
for the term 'resource' is undefined in the earliest specification for “Universal
Resource Identifiers” (Berners-Lee 1994a). Berners-Lee has remarked that one
of the best things about resources is that for so long he never had to define
them (Berners-Lee 2000). Eventually Berners-Lee attempted to define a resource
as “anything that has an identity” (Berners-Lee et al. 1998). Other specifications
were slightly more detailed, with Roy Fielding, one of the editors of HTTP, defining
(apparently without the notice of Berners-Lee) a resource as “a network data object
or service” (Fielding et al. 1999). However, at some later point Berners-Lee decided
to generalize this notion, and in some of his later works on defining this slippery
notion of 'resource,' Berners-Lee was careful not to define a resource only as infor-
mation that is accessible via the Web, since not only may resources be “electronic
documents” and “images” but also “not all resources are network retrievable; e.g.,
human beings, corporations, and bound topics in a library” (Berners-Lee et al.
1998). Also, resources do not have to be singular but can be a “collection of other
resources” (Berners-Lee et al. 1998).
Resources are not only concrete messages or sets of possible messages at a given
temporal junction, but are a looser category that includes individuals changing over
time, as “resources are further carefully defined to be information that may change
over time, such as a service for today's weather report for Los Angeles” (Berners-
Lee et al. 1998). Obviously, a web-page with “today's weather report” is going
to change its content over time, so what is it that unites the notion of a resource
over time? The URI specification defines this tentatively as a “conceptual mapping”
(presumably located in the head of an individual creating the representations for