Problems with metadata management plague organizations of all sizes, industries, and geographical locations. Information-rich industries are especially prone to experiencing metadata-related pain
and would therefore be wise to consider the relevant technological and process solutions.
A leading role in most metadata management solutions is played by metadata repositories, be they home-grown internal creations or vendor tools. Metadata repositories generally support the assorted
data aspects of IT systems and include both Technical and Business Metadata. Technical Metadata generally supports IT staff and users, while Business Metadata supports an organization’s business
users.
Metadata repositories go beyond the IT staff, to a wider enterprise user base, in providing knowledge management strategies across organizations. According to David Marco, author of “Building and
Managing the Meta Data Repository: A Full Lifecycle Guide:” “I don’t see how a ‘true,’ enterprise wide knowledge management solution can exist without a meta data repository. In fact, the meta
data repository is the backbone of a knowledge management solution.” (source: David Marco, “A Meta-Data Repository is the Key to Knowledge Management”).
Nonetheless, the growth of metadata repository implementations (“the number of repositories used in the industry will double,” source: Michael Blechar, “Magic Quadrant for Metadata Repositories,
2004,” Gartner Research Note) in recent years has resulted in difficulties not originally perceived by most project leaders and organizations. Some examples of the challenges facing metadata
repository projects now and in the future are the need to expand support for different types of information, the need to bridge between technical and business users, and the need to simplify
configurability. Without resolving these and other issues, there exists a strong likelihood that many metadata repository implementations will fail or only partially succeed.
Semantics define a concept’s meaning in a manner that is both unambiguous and universally correct in meaning. Precise semantics ensure that an organization’s conflicting definitions and
terminology are accurately understood and also distinguishable by the enterprise’s people or software. While semantics affect our everyday life as individuals (e.g. synonyms or linguistic
differences between technical jargon and popular speech), they are just beginning to be applied to IT environments. Applying the precise expressiveness of semantics to IT generally, and metadata
specifically, enables applications, systems, data sources, and personnel to relate to and to understand concepts in a consistent fashion.
Semantics is being applied to metadata management issues as well. Semantic metadata repositories are catalogs of metadata in which the precise meaning of each term is expressly defined. As a
result, semantic metadata repositories share the benefits of traditional metadata repositories and also bring additional advantages to metadata projects.
For instance, semantic metadata ensures that technical and business users are relying on common business meaning, regardless of how it is represented or referred to. This reduces the all too
frequent communications gap that exists within large organizations between IT and the business. By promoting the use of metadata which is consistent for each of these parties and also correlated to
one another, semantic metadata management enables users to continue using the various data management systems and life cycle management they are used to, without disrupting their ongoing work. In
point of fact, semantic metadata repositories are also relevant to resolving communication differences solely existing within IT teams and among business users.
Semantic metadata repositories enable the use of consistent data by rationalizing all physical data sources; this is accomplished by mapping each data source to an ontology. Ontologies are highly
expressive models of specific domains of knowledge. Originating in Greek philosophy and until recently a focus of computer science departments, ontologies have proven to be robust creatures
attracting widespread interest. For instance, the W3C recently published its Candidate Recommendation of the Web Ontology Language (http://www.w3.org/2001/sw/WebOnt) in February 2004 as part of its Semantic Web Activity (http://www.w3.org/2001/sw).
The Semantic Web idea that data has objective meaning – enabling it to be dynamically found and used by computers – enables new systems being removed or created to be discovered as they are
modified or come into existence. According to the W3C, “The Semantic Web provides a common framework that allows data to be shared and reused across application, enterprise, and community
boundaries. [It] is an extension of the current web in which information is given well-defined meaning, better enabling computers and people to work in cooperation (http://www.w3.org/2001/sw).”
The combination of semantics and traditional metadata repositories resolves a number of hitherto difficult problems which have disrupted numerous metadata management initiatives. For instance, the
“variety of metadata involved in defining the mapping of the data” needs to be accounted for in most organizations; this is a benefit of semantic metadata repositories, a trait which traditional
metadata repositories generally lack. (source: Malcolm Chisholm, “Is the Meta Data Repository Dead?”; http://www.dmreview.com/article_sub.cfm?articleId=3335). In addition, most organizations requiring metadata repositories need to achieve
significant levels of configurability, scalability, and openness towards future endeavors and possibilities. While not providing a foolproof guarantee of such benefits, applying semantics to
traditional metadata cataloging efforts does provide an enormous degree of assistance to the users of such tools.
As with any IT initiative, the use of semantics with metadata repositories entails certain risks. One of the biggest possible problems regarding semantic metadata repository initiatives has to do
with human nature. While resolution of conflicting definitions and meaning is crucial, often commitment to reaching such resolution is missing. The result is incorrect and partial semantic
expression! Moreover, the need to maintain such a system is important to both creators and users of such projects. As with most IT projects that attempt to sit astride more than one team or
department, metadata cataloging schemes are ideal candidates for abandonment, a state of affairs which is likely to cause confusion with both current and future users relying on such systems. So
while this last point is not a problem unique to semantic metadata repositories, the additional effort of applying semantics to metadata repository construction results in an extra layer of
complexity.
The combination of semantics and metadata repositories makes possible the resolution of some of the most pressing issues within the world of metadata management. This will be a boon to
data-intensive organizations, specifically those advancing their metadata repositories. In addition, the convergence of this topic with the ambitious Semantic Web effort headed by Tim Berners-Lee
and the W3C surrounding ontologies, semantic markup, and semantic interoperability may just be enough to bring semantic metadata management into the forefront of mainstream IT.
