Data Architects are fond of criticizing clients for not properly handling language. Throughout the enterprise the same word is used to mean different things. Different words are used to mean (approximately!) the same thing. An important part of any data governance project is to help the client come to terms with the way it uses language.
But what about us in the data management profession? If a company advertises for a “data modeler”, what exactly is it expecting? For that matter, if you answer the ad, what are you prepared to do?
How do you define a “conceptual model”? What about a “logical model”? “Physical model”? Would your neighbor agree on any of these? But of course, that’s not all. We have to deal also with “entity relationship model”, “object role modeling”, “the semantic web”, and “XML Schema”, among others. Oh, and can anyone define “canonical model”?
It turns out that the data industry is beginning to converge on these and related terms. But it’s not quite done yet. This presentation is a humble attempt to complete the process.
Based on the industry’s history, his proficiency in language and semantics, and, OK, some personal prejudice, David Hay will present (at least his version of) a definitive vocabulary for our field. This will cover not only the traditional terms “Conceptual”, ‘Logical”, and “Physical” models, but introduce two new ones, just to make things interesting: “Semantic” and “Essential” models.
In each case, the term will be coherently defined, along with examples of the techniques appropriate for each.
A moderate amount of audience participation is welcome—as long as the objective is to refine the definitions, not simply to disagree with them.