Just how important is the name we call something? In some cases, it is extremely important, but in other cases it isn't. To separate the two, it is important to understand why we attach a name to things and people. The primary reason for naming something is to facilitate communication. I recently attended a party at which I met a person whom I had not seen for several years. While I recognized the face right away, I did not recall the name. As we were talking, I remembered other facts about the person. Not knowing the person's name did not inhibit our conversation. When my wife joined me and I was faced with making introductions, not knowing the name became a problem because it was needed for communication. Prior to that time, it was the information about the person that was of greater significance.
The same is true in data warehousing. As we read articles in the press, we are bombarded with terms such as data warehouse, operational data store, data mart, OLAP, ROLAP, MOLAP, decision support, meta data, etc. It is also not uncommon for several articles to use the same term but to mean completely different things by it. The terms are used for communication purposes. Just as the same word can have multiple definitions, in data warehousing the same term may be used to mean different things. While there are sources, such as unabridged dictionaries, that provide the alternate meanings for English words, no such compendium provides a comprehensive set of variations for information technology terms. Therefore, to ensure that the communication is complete and accurate, the meanings must be used (or implied by reference) in conjunction with the terms.
Let's look at the operational data store (ODS) as an example. A common definition for an operational data store is that it is a subject-oriented, integrated, current, volatile collection of data used to support tactical decisions.1 An ODS has also recently been defined as the front edge of the data warehouse.2 The two definitions have a few elements in common--both recognize that the ODS contains integrated, subject-oriented data. They differ, however, with respect to the role of the ODS, in the way the ODS is updated and in the data that it may contain. The differences do not invalidate either definition. Since each of the definitions is being professed by one or more widely recognized authorities in data warehousing, the term operational data store has ambiguity attached to it if used without being qualified.
Just as each data element needs to be defined to facilitate communication,3 each construct being developed also needs to be described. A company implementing what it terms an ODS must define the term for internal and external communications purposes. Internally, the definition ensures that the people working on a project and the business people with a need to be satisfied understand the business issues. It also governs the project approach and its deliverables. Externally, the definition helps readers of articles, such as those noted, to appropriately glean the useful information so that they can leverage others' experiences and knowledge. The definition also helps in communications with other companies so that their experiences can also be appropriately evaluated.
I view a data mart as being a derivative of the data warehouse, structured to meet specific business unit or functional needs. When I am asked to submit a proposal, however, my definition is not the key factor. Before I can respond to a company's request, I must determine what the company means by the terms it is using. Sometimes, what a company refers to as a data mart may be what I consider to be a data warehouse. A data warehouse for some is an operational data store for others. This does not necessarily mean that the company's representative is in error--it means that the company's chosen definition does not agree with mine and that before I can provide a reasonable proposal, I need to understand the company's needs, independent of the terminology.
The name attributed to an architectural construct such as a data warehouse, operational data store or data mart is important, but it is not as important as the understanding of what it takes to build and maintain the construct. This understanding is critical for a company to determine the methodology and tools that best apply. The name attributed to the data store is needed for communication. It is the combination of the name and definition that enables companies to successfully share information and enable readers of articles to relate information from the articles to their environment.
1 Battas, Greg. Imhoff, Claudia. Inmon, W.H. Building the Operational Data Store. John Wiley & Sons, 1996.
2 Kimball, Ralph. "Relocating the ODS." DBMS. December 1997. P. 12.
3 Geiger, Jonathan G. "Data Element Definition." DM Review. December 1996. P. 38.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access