Quality measures for and Transformation of conceptual schemas
In the 1980ies conceptual schemas are recognized and accepted as an important tool for the design and evolution of integrated databases and knowledge-based systems. But the question of quality of conceptual schemas is largely ignored by research. In relational database theory quality is defined by the presence or absence of certain normal forms. The definition of quality is very restrictive because a conceptual schema is either good or bad. Other criteria such as the complexity of a schema are not important. Quality measures are also ignored. But transformation of conceptual schemas are explored systematically.
So a number of criteria are introduced to ensure that transformations are valid to the original schema. The lack of quality measures make the use of these transformations not very useful because there is no clear understanding of a good conceptual schema.
Christoph F. Eick represents in his paper “A Methodology for the Design and Transformation of Conceptual Schemas” the back end of a conceptual schema design methodology, called ANNAPURNA. This methodology aims to automate conceptual schema design focussing on the transformation and evaluation of conceptual schemas based on quality measures and transformations that has a theoretical foundation.
Quality measures for conceptual schemas are introduced, a general framework for the specification of conceptual schema transformations is proposed and algorithms for evaluation and transformations are provided.
So Eick is one of the first who introduces quality measures for conceptual schemas and definitions of a valid and good schema.
Eick also says that it is not desirable to use completely automatic design tools without human assistance because there are several reasons why the cooperation between a human expert and a design tool is still necessary. Such reasons are the acquisition of dependencies by asking human experts and that the finding of the best or even only a good conceptual schema may require a lot of heuristic knowledge to be automatically feasible.