An XML-based implementation of the parametric model for ad-hoc query of temporal and spatiotemporal data

The parametric model is one of the data models for dimensional data. Values in the parametric model are defined as functions. Such modeling concept helps one achieve a one-to-one correspondence between objects in the real world and records in a database. One of the important requirements is that domains of values should be closed under the set theoretic operations such as union, intersection, and complementation. Because of this, ParaSQL, a query language of the parametric model, is able to mimic natural languages more closely. In this dissertation we validate and implement the parametric model for temporal and spatiotemporal data. We also develop a preliminary prototype for the users of NC-94, an interesting dataset in agriculture.; Viewing values as functions leads variable-length tuples. Potentially, such values vary in size ranging from a few bytes to gigabytes and beyond. This makes implementation of the parametric model a challenging problem. To meet the challenge, we develop an XML-based storage and deploy it in our implementation. Incidentally, XML is also used for interfacing various modules and artifacts like parse tree, expression tree, and iterators to fetch data from a disk.; The NC-94 dataset, mentioned above, contains the most complete record of spatiotemporal variables that characterize the dynamics of agriculture covering the north central region in the United States. To support ad-hoc query of data in its geospatial context, a novel hybrid structure is designed and implemented. We use GML to describe geospatial information. Use of GML is a good match, because it is XML-based. More importantly, it meets the set theoretic closure requirements proposed by the parametric model.; Validation and implementation methodologies introduced in this dissertation will contribute to database and GIS communities. The validation demonstrates the ease of use and efficiency of the parametric model for temporal and spatiotemporal data. This should help settle a debate in temporal database community which has continued since the mid 1980s. The findings also extend to spatial and spatiotemporal data. It is an important baby-step toward full-fledged implementation of the parametric model. We hope that this work will also help bring database and GIS communities together.