In my last set of posts I started to look at some of the challenges associated with enterprise management of reference data domains, especially as the scope of use for the same conceptual reference domains expands across databases, systems, and functional areas within the organizations. Recognizing the value of capturing business metadata about reference data provides two benefits.
First, collecting metadata about reference domains provides a less-demanding way of jump-starting a general metadata process as well as identifying the basic requirements for a tool or management facility to support a metadata management initiative. Structural metadata for reference domains describes the data type and structure of the values, which in many cases (such as ISO-3166 2-character country codes) are constrained to a particular length and format.
However, the qualitative “information” about reference data represents not just the name of the reference domain and the values it contains. It also include the core concepts being represented by the domain values, the commonalities that bind them together within the domain, and the interpretations and meanings associated with each value in the context of the domain. These are all valuable facets of information to capture within a metadata repository, and it paves the way for expanding the use to data element concepts, data elements, and collections of data elements as tables or relations.
The second benefit? Mapping the instances of shared reference data to their uses across the organization allows data practitioners to accumulate semantic knowledge about perceived definitions of common domains. These become a model for more complex shared data domains. After looking at some of the complexities of implementing a master data management program, I am confident that some of the challenges stem from not understanding the context of shared information in a distributed environment.
A lot of the protocols for managing master data are integral to managing shared reference data. In fact, managing reference data as a shared data resource and developing methods for publishing the definitions, structures, and semantics of reference tables as well as developing methods for shared accessibility provide models for more complex master data domains.