Architecture of theWorld WideWeb - Social Semantics: The Search for Meaning on the Web - page 27

Fig. 2.4 An example URI, with components labelled

place in the path component and '..' as the previous level in the path component. So,
../EiffelTower is a perfectly acceptable relative URI. Relative URIs have a
straightforward translation into absolute URIs, and it is trivial to compare absolute
URIs for equality (Berners-Lee et al. 2005).
The 'hash' ( # ) and 'question mark' ( ? ) are special characters at the end of
a URI. The question mark denotes 'query string.' The 'query string' allows for
the parameterization of the HTTP request, typically in the cases where the HTTP
response is created dynamically in response to specifics in the HTTP request. The
'hash' traditionally declares a fragment identifier ,which identifies fragments of a
hypertext document but according to the TAG, it can also identify a “secondary
resource,” which is defined as “some portion or subset of the primary resource, some
view on representations of the primary resource, or some other resource defined or
described by those representations” where the “primary resource” is the resource
identified by the URI without reference to either a hash or question mark (Jacobs
and Walsh 2004). The fragment identifier (specified by a 'hash' followed by some
string of characters) is stripped off for the request to the server, and handled on the
client side. Often the fragment identifier causes the local client to go to a particular
part of the accessed HTTP entity. If there was a web-page about Gustave Eiffel, its
introductory paragraph could be identified with the URI http://www.example.com/
EiffelTower#intro . Figure 2.4 examines a sample URI, http://example.org/Gustave
Eiffel#birthday :
The first feature of URIs, the most noticeable in comparison to IP addresses,
is that they can be human-readable, although they do not have to be. As an idiom
goes, URIs can be 'written on the side of a bus.' URIs can then have an interpretation
due to their use of terms from natural language, such as http://www.whitehouse.gov
referring to the White House or the entire executive branch of the United States
government. Yet it is considered by the W3C TAG to be bad practice for any agent
to depend on whatever information they can glean from the URI itself, since to a
machine the natural language terms used by the URI have no interpretation. For an
agent, all URIs are opaque, with each URI being just a string of characters that can
be used to either refer to or access information, and so syntactically it can only be
checked for equality with other URIs and nothing more. This is captured well by
the good practice of URI opacity , which states that “agents making use of URIs