Over the past few years there has been an emergence of registries due to the lack of organization of the World Wide Web. These registries provide a mechanism to centrally organize information in a way that makes it easier for both people and machines to locate the required sources of information.

However, there is also an alarming new trend taking place now, which is the proliferation of registries, each with a different set of data structures, access mechanisms, and expected functionality. At present there are many parallel registry standards, such as UDDI (Universal Description, Discovery, and Integration), ebXML, .NET Passport, Domain Names Services (DNS), and so on. While these standards aren't necessarily exclusive of one another, each has a different function and audience it intends to satisfy.

I propose that these standards, while seemingly useful relative to their tasks, can all be subordinated to a properly designed Semantic Web initiative. The outcome of such a venture would produce a single methodology for the storage and retrieval of any type of registry information as well as facilitate reusability across registry types. Additionally, highly focused programming interfaces wouldn't be necessary as the single, abstract registry model would support the needs of creating and querying specialized subsets of a larger registry model.

What Is a Registry?
One of the key attributes of a registry is metadata, which also happens to be one of the key factors of XML that has endeared it to the IT community. A registry is a centralized store that contains individual entries, each representing information (metadata) pertaining to a particular category.

UDDI is an example of a registry for Web services. The registry contains all the information necessary to contact the provider of the Web service and how to bind to the service from an application. Providing this consistent set of information across all entries allows the registry to then provide query facilities to make the registry more useful. However, because UDDI is a focused registry, you won't find information on how to learn Web services or Web service standards.

The UDDI registry also defines specific XML structures that must be understood by any user of the UDDI registry, such as the businessEntity, businessService, and tModel. Each of these structures is defined by the UDDI organization and is proprietary to that registry. That is, they can't be used effectively with other registries, such as ebXML.

The Semantic Web Initiative
The Semantic Web, a W3C Activity (www.w3.org/2001/sw) led by Tim Berners-Lee, James Hendler, and Ora Lassila, is defined by the following quotation: "The Semantic Web is an extension of the current web in which information is given well-defined meaning, better enabling computers and people to work in cooperation."

One of the key enabling technologies for the Semantic Web is the Resource Description Framework (RDF), which is another initiative within the Semantic Web Activity within the W3C. RDF has an associated XML syntax for expressing a directed graph of nodes connected by arcs denoted by Universal Resource Identifiers (URI). Listing 1 is an example of an RDF taken from the W3C RDF/XML Syntax Specification Working Draft document that describes the relationship between a Web page and its publisher.

As you can see from the listing, the RDF document creates an association between the editor (Dave Beckett) of the document located at www.w3.org/TR/rdf-syntax-grammar and his homepage (http://purl.org/net/dajobe/). In this way the RDF can create flexible associations between disparate pieces of information.

How the Semantic Web Can Influence Registry Development
The RDF in Listing 2 could also represent the same information set found in the UDDI registry.

In the previous example, what I hoped to prove is that we already have an effective framework for the categorization and association of information identified with the Web infrastructure in RDF. With this framework in place, what we need is a consistent programming interface to allow this information to be managed, replicated, and updated in a consistent fashion. Creation of individual and distinct registries, each with their own structures, query facilities, and programming interfaces, is unnecessary and will lead to downstream hurdles for interoperability, reusability, and heightened restrictions on usage.

Additionally, Query, which seems to be one of the more predominant and interesting tasks being levied against these registries, can be far more useful and less proprietary when an abstract registry representation such as RDF is used. That is, only one search engine would be needed to query the arcs and nodes of an RDF document, and an RDF document could represent, in an equivalent manner, a UDDI business service and an ebXML Collaborative Protocol Agreement.

Conclusion
The resulting work of the Semantic Web Activity can be used to create a global registry framework in which information can be identified by multiple associations without creating barriers to entry, and to provide a single methodology for search and retrieval of information within this network of information.

Likewise, when coupled with standards for creation of ontologies (www.w3.org/2001/sw/WebOnt), the Semantic Web offers the ability to create multiple views of the same set of relationships in varied forms so as to best fit the needs of the user community without the need for specialized registries.