27.1 Ontology and Schema Fragility

One of the major problems in computer science is the question of how best to represent or organize data, for example by a data structure or a database schema. Indeed, volumes have been written on data modeling, and numerous methodologies have been developed to codify this task; object-oriented design, relational design, object-relational design, and so on. Two measures by which these design-styles can be evaluated are (i) how well they can present the same information in multiple views or organizations and (ii) the related problem of how well they can compensate for changes or modifications to their schema/organization.

These two challenges form the basis of ``Schema Fragility''. Different users or applications may require the same data to be organized differently for ease of processing, or reflecting different models or levels of detail, or to provide different user communities with views customized for their perspective. This ability is not only critical to the ``applicability'' of a representation, but also its ``longevity''. One of the most important aspects of modeling a ``real-world'' system is that both it and our knowledge of it are changing continually. A robust design is one that can easily adapt to such inevitable changes.


Subsections