Upto: Table of Contents of full book

Uniform resource Identfiers

Objects need identifiers. Concepts such as relationships also need identifiers. These need to label things uniquely. This role is played by URIs (Uniform resource Identfiers)

Resources

First there were URLs...

Hyperlinks formed a critical part of the World Wide Web from the start. They gave a simple mechanism for pointing from one Web page to another. They were known as URLs (Uniform Resource Locators). However, they were made up of several parts

The user agent needs to know how to fetch a linked resource. This could be by FTP, HTTP or by some other network protocol. This part of a URL is called the scheme and is for the HTTP protocol just http
Web resources reside on servers, that is computers on the internet. The server's DNS (domain name service) address must be known. These are the names we are familiar with, such as www.google.com
Each server will have its name for resources. These appear after the server name
Fragments within an HTML document will sometimes be labelled, after a #

A typical URL, e.g. for this page is

http://jan.newmarch.name/SemanticWebForLibraries/URIs/index.html

where http is the scheme, jan.newmarch.name is the addres of my server and SemanticWebForLibraries/URIs/index.html is the name of this resource on my server.

Then there were URNs

But nobody ever used them, so (almost) forget about them. Except they hang around in some places. They begin with urn

And now there are URIs

URLs in a crude sense label where things are. (Crude, because you don't really know where e.g. Microsoft servers are.). But if you wanted to label everything, then, well, not everything has a location. For example, ISBN numbers aren't found at a particular place. Even worse, relationships aren't found anywhere: if "father-of" referred to a particular place then there wold be many paternity suits...

URIs (Uniform Resource Identifiers) are the Web solution to naming things. The basic structure is fairly vague:

scheme : something

The reason for the vagueness is to accomodate as many schemes as possible. For example, http schemes are very precise for labelling where (or which parts of) a resource are located. Other schemes include the ISBN scheme, a Z39.50 Retrieval Scheme, a Z39.50 Session scheme, and many hundreds of others.

Who decides what?

The format of ISBN numbers was established as an ISO standard. How does this and other schemes relate to the Web? Well, it's a complex story!

Web standards are maintained by the W3C, the World Wide Web Consortium. They set the standards for HTTP, HTMl, XML, RDF and a huge variety of other standards
General oversight of the internet is managed by the IETF, the Internet Engineering Task Force. Where appropriate, they adopt the standards set by the W3C. So HTTP is IETF RFC 2616 because that is a protocol used on the internet. HTML is not considered by the IETF because it is all about content, not transport.
Schemes are strings (with meaning) that have to be recognised world-wide. As far as the internet is concerned, IANA (Internet Assigned Numbers Authority) is the authority for these. It manages the possible scheme names at Uniform Resource Identifier (URI) Schemes
Now this doesn't actually include the scheme for ISBN numbers: this is one of the hangovers from the days of URNs. These are also looked after by IANA, but at a different place Uniform Resource Names (URN) Namespaces
The IANA scheme for ISBN points back to an IETF RFC Using International Standard Book Numbers as Uniform Resource Names which finally refers to NISO/ANSI/ISO 2108:1992 as the authority for ISBN numbers

At the end of this chase, you would come up with the following for ISBN numbers:

URN:ISBN:<:ISBN number>

In a similar way you can find URI encoding schemes for other labelling schemes in existence.

Conclusion

Objects, ideas, relationships need to be labeled in some way. URIs are the Web way of labelling things. While a significant subset is URLs, other shemes such as ISBN can be brought into the URI world. As we shall see later, URIs play a fundamental role in all aspects of the Semantic Web.

[an error occurred while processing this directive]