Following our review of the Data Management Body of Knowledge (DM-BOK) last month, we ordered the companion volume—the Dictionary1, sold as a CD Rom containing a single PDF file. We were curious to see if this work would be the Linnaean classification of data management that we have been looking for—and, more prosaically, how its definitions stack up with Wikipedia. First impression is that it is rather a lightweight production, the 850 definitions occupy 140 pages of rather spaced-out text, probably not more than a 30,000 word count in the whole dictionary.
In the introduction, the authors note—and we can only concur, that the industry is ‘in great need of clarity in its terminology.’ They further observe that ‘While experts may never reach 100% agreement, we believe there is general agreement with these definitions across the data management profession.’ Well that would be nice if it were true, but one of Wikipedia’s strengths is the way it captures enlightening differences of opinion.
So what’s it worth? The definition for ‘online analytical processing’ is ‘The use of multi-dimensional databases and/or analytical tools to support the analysis of business trends and development of business projections.’ This is OK as far as it goes—but it lacks narrative. It fails to explain the relationship between a normalized data base’s poor query performance. Wikipedia hits the nail on the head in its first sentence viz., ‘an approach to swiftly answer multi-dimensional analytical queries.’
How about ‘master data?’ This is ‘synonymous with reference data’ and is ‘data that provides the context for transactional data including details of internal and external objects involved in business transactions.’ That is OK but not much use outside of the business/financial services environment of DAMA.
The synonym, ‘reference data’ curiously has a slightly different definition ‘data used to categorize other data, or for relating data to information beyond the boundaries of the enterprise—see master data.’
Semantic ‘has to do with meaning, usually of words and/or symbols’ while a semantic data model (synonymous with an ontology) is ‘a [..] data model for non-tabular data [that] makes meaning explicit so that a human or software agent can reason about it.’ Another lexical run-around occurs when you check out ontology—‘a semantic data model [as above]’ and ‘a synonym for schema.’
I am afraid that the DAMA Dictionary’s usefulness is rather limited. It might be useful for looking up the odd acronym or checking out a business-oriented definition of a term? But Wikipedia beats it hands down for depth and, we believe, authority. Wikipedia, unlike the Dictionary’s CD-ROM/PDF format is also more likely to be to hand when you need it.
1 The Data Dictionary of Data Management 1st ed. 2008, Mark Mosley, Editor, Technics Publications. 140 pages. $44.95.
This article originally appeared in Oil IT Journal 2010 Issue # 7.
For more information or to comment on this topic email here.