I read Data and Reality by William Kent in order to gain a diiferent perspective on this difference between the database model and reality.
In describing the system uses a database, Kent (2000, 28) comes up with the three (3) level model of description which corresponds to the ANSI model:
- “The multiple views held by a variety of applications, each employing their own variations on record formats, structures, and access techniques. This level is variously referred to as “user”, “application”, “external”, “program”, and “logical”.
- “The physical layout of data in storage, including implementation techniques for various paths and linkages. The common names for this level are “internal”, “storage”, and “physical”.
- “The specification of the information content of the database, employing concepts equivalent to enities, attributes, and relationships. Names for this level include “conceptual”, “information”, and “entity” (and, sometimes, “logical”)”
So far, I have encountered several authors who use this three (3) level model of the database modelling process. The ANSI model is now a consensus. The only exceptions have been Churcher (2007) and Hernandez. I would prefer to keep the three (3) level model because of its robustness.
Kent describes a model as follows:
A model is a basic system of constructs used in describing reality. It reflects a person’s deepest assumptions regarding the elementary essence of things. It may be called a world view. (Kent 2000, 107)
I think Kent makes a mistake in assuming that reality is independent of the observer. As human beings, we can only sense things that our senses allow us to. Being human constrains our perceptions of the external world. I will not go into any philosophical discussions about whether the external world exists.
Within those perceptions, only a subset of those are culturally important. We do not react to all sense perceptions. Those that we do are important because we assume that they are important in our culture. The facial expression of a guy with a gun is far more important than the angle of sunlight on a leaf. The former is more culturally significant in that our survival depends on the correct interpretation.
In other words, what various authors call the real world is a human cultural construct. The problem for database designers is to convert this construct into a computer construct.
Later on, Kent makes a very important point:
Computers do deal only in signs, and a database is`only a collection of signs. From these premoses one can easily — and incorrectly — conclude that the conceptual model describes a collection of signs. (Kent 2000, 112) (Emphasis Mine)
My reading of Kent (2000, 113) is that a conceptual model is the bridge between the set of things and the set of signs. He says that:
…we can more clearly approach a semantic bridge to reality in the conceptual model, by explicitly relating sets of signs to sets of things. (kent 2000, 113)
The objective for Kent would appear to be the construction of this semantic bridge.
Kent, William (2000), Data and Reality, 1stBooks : USA.