Biophysics of computation in single hippocampal neurons

Thomas Huntington Brown, Anthony M. Zador · 1993

A major application area of object-oriented databases is for design systems. Existing object-oriented data models surpass traditional models in modeling the complex structures and intricate relationships of design entities. However, they cannot address sufficiently the dynamic evolution and flexible versioning needs of actual design objects. This thesis presents a new object-oriented data model, called the object-centered data model, that satisfies the modeling requirements of design databases. In this model, objects are represented as independent structural entities and classes as structural constraints enforced for class membership. Objects can be dynamically associated with classes to inherit their behavioral definitions. It is shown that such a loose and non-permanent binding between objects and classes allows objects to evolve both in their structure and behavior during the design process. In spite of the dynamic nature of objects, strong typing can still be achieved by enforcing a minimum set of evolution rules. I also show that the model supports multiple and changing object classifications due to its capability for defining many separate class hierarchies in a database. Based on the object-centered data model, a version model supporting both object and schema versioning is also designed. In object versioning, the version model uses generic objects to group object versions of different structure and behavior so that the complete spectrum of the evolution process of design entities can be captured. In schema versioning, the model uses individual class hierarchy as the granularity of schema versioning. It greatly reduces the cost for sharing instances among different schema versions by considering schema versions as different views of their instances. For configuration specification of complex objects and classes, the model supports several forms of dynamic version references and a sophisticated context mapping mechanism. Finally, it incorporates a mechanism for enforcing a variety of versioning-related constraints to assure consistent evolution of versioned entities.

Read the paper · More papers on PaperTik