Metadata: A fundamental component of the Semantic Web

Bulletin of the American Society for Information Science & Technology, Apr/May 2003 by Greenberg, Jane, Sutton, Stuart, Campbell, D Grant

Proof and Trust. The last two horizontal layers build off of the logic layer, proof being a validation of the "evidence" stemming from the inferential logic activity and trust relating to the integrity of the proof, which may be traced back down through the other layers in Berners-Lee's diagram. The functionality of these two layers is highly dependent on creation of accurate and trustworthy metadata. Digital signature. Digital signatures run horizontal to the RDF family up through the proof layer and support the notion of trust. Developments in the area of digital signatures are progressing, and could eventually help validate the integrity of metadata that an agent will use for reasoning and task completion.

The synopsis provided here shows that metadata permeates each layer of the Semantic Web's architecture, although it is not the only piece, as agents, enabling technologies, authoring and annotation programs, and metadata vocabularies need to be further developed to realize the full potential of the Semantic Web.

Metadata Vocabularies

Metadata vocabularies are synonymous with onotologies, as discussed above. A vocabulary, in a general sense, is a shared system of semantics for concepts or objects. Think about the vocabulary that comprises Portuguese. It is a system of agreed upon meanings permitting intelligible communication among Portuguese people and other persons who speak this language.

The metadata world hosts a range of systems to which the label of "metadata vocabulary" is applied. These vocabularies range from basic descriptive metadata systems, with limited or single-focused functionalities, to more complex "member/ class" semantic vocabularies. An example of a simple ontology is the Dublin Core Metadata Initiative (DCMI) Elements and Element Refinements (www.dublincore.org/usage/ terms/dc/current-elements/), a metadata schema developed mainly to facilitate resource discovery. A more complex metadata vocabulary to which the word ontology is applied is the Ariadne Genomics ontology (www.ariadnegenomics.com/ technology/ontology.html). This ontology is used to formalize data on cell proteins for computer analysis: the ontology defines the various proteins, classifies them into taxonomic trees and defines the semantic relationships among them. While many metadata vocabularies in operation were not necessarily created with the Semantic Web in mind, they may be able to play a significant role in its development. For existing and developing ontologies to be used and function fully in the Semantic Web environment, they need to adhere to standards supported by enabling technologies.

Enabling Technologies for Metadata

Although metadata is integral to the Semantic Web, metadata on its own is far from sufficient. Needed are standards to syntactically encode and represent semantic knowledge so that agents can perform tasks in an efficient and comprehensible manner. A variety of enabling technologies have developed over the last few years that are proving significant to the construction of the Semantic Web. Some have developed prior to the conceptualization of the Semantic Web, while others are being developed and refined specifically with the Semantic Web in mind. Among four key developments that are critical for metadata encoding and manipulation by agents are: