Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
2000-10-06
2003-10-28
Vu, Kim (Department: 2172)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000, C706S047000
Reexamination Certificate
active
06640231
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention generally relates to database systems and, more particularly, to a process for creating and maintaining ontologies and processes for semi-automatically generating deductive databases, especially for biological information systems.
2. Background Description
Database engineering practices and technologies of the last two decades have proven a poor match for the complex information handling and integration needs of modern enterprises. These techniques and technologies center on the construction of databases which are manually designed by a team of developers. This not only includes the database code itself (schemas, views, and integrity constraints) but also includes the peripheral software needed to run such a system: data loaders or “cleaners”, application software, and other computer resources.
For very simple or highly standardized domains, this approach is sufficient. Simple domains require a simple database schema and few integrity constraints on the data. Such domains change very slowly over time, making it easy for developers to keep up with the design requirements. However, many problems faced by modern database customers don't fit these criteria. For instance, systems involving any sort of analytic component typically require extremely complex and fluctuating rules reflecting real-world situations. Using current techniques, the process of keeping such information systems current is error-prone and prohibitively expensive, costing millions of dollars in developer salaries alone over the system life cycle.
Moreover, current systems have a fundamental and more severe problem: integrating data from any two systems requires custom-made middleware, because it is impossible for the system to “understand” the content of the participating databases well enough to perform the required integration automatically. The use of a shared ontology to enable semantic interoperability of existing databases and other software is gaining acceptance. It is possible to enable communications between two systems by mapping the semantics of independently developed components to concepts in an ontology. In the computer sciences, an “ontology” refers to a conceptual model describing the things in some application domain (e.g., chemistry) encoded in a formal, mathematical language.
In the context of the invention, an ontology is a formal (concretely specified) description of a business domain. It contains a taxonomy of concepts (“a person is a type of mammal”; “a corporation is a type of legal entity”), and also contains a set of rules relating those concepts to each other (“flight numbers are unique within airlines over time”). Data element standards and metadata repositories and their associated tools formalize some (but not all) system behavior, leaving the rest to be specified in free-form English text which cannot be “understood” automatically. Ontologies, on the other hand, represent these concepts and rules in a completely formal language; their meanings are meant to be accessible to the computer. Unfortunately, ontologies are specified using languages which are far too powerful to allow their being used in a straightforward manner to build practical information systems, until development of the present technology.
Generating application-focused databases from large ontologies is described by Brian J. Peterson, William A. Anderson and Joshua Engel in
Knowledge Bus: Generating Application
-
focused Databases from Large Ontologies
, Proceedings of the 5
th
KRDB Workshop, May 1998 (hereinafter, Peterson et al.) and herein incorporated by reference in its entirety. In their paper, Peterson et al. propose to generate the databases (including application program interfaces (APIs)) directly from focused subsets of a large, general purpose ontology. By extracting only a subset of the ontology needed to support representation and reasoning in a focused application domain, the resulting systems are smaller, more efficient and manageable than if the entire ontology were present in each system.
SUMMARY OF THE INVENTION
The subject invention builds on the work of Peterson et al. According to the invention, there is provided a process for creating and maintaining ontologies and a process for semi-automatically generating deductive databases (DDBs). The ontology is a Ontology Works language (OWL) ontology managed by the Ontology Management System (OMS). An OMS ontology has a hierarchy of categories, which denote classes of objects (note that this is different from the object-oriented notion of class). This hierarchy is partitioned by the type and attribute hierarchies. The type hierarchy includes the categories that can participate in predicate signatures, and corresponds to symbols that become types within a generated database. The OMS ontology consists of a set of OWL sentences, each of which has an associated conjunctive normal form (CNF) version. The deductive database generator (DDBG) applies a series of conversion and review steps on the CNF of the OWL sentences within the input ontology. It generates a pre-DDB, which defines the schema of the deductive database, as well as provides the rules required for reasoning the integrity-constraint checks. A Strongly-Typed API Generator (STAG) takes the pre-DDB and generates a Java-based API for the resulting DDB. This API is a strongly typed, object-oriented view of the elements defined in the pre-DDB. The DDB consists of a pre-DDB with a Java server and a backing store.
The following sections describe the process used to generate databases:
Extraction
The extraction phase starts with those entities and relationships immediately relevant to the problem at hand, and identifies those parts of the ontology necessary to support them. For example, a dinosaur taxonomy is not relevant to a database supporting financial analysis of automobile exports, but concepts relating to products and international economics are. The set of immediately relevant concepts may be already present in the ontology, entered by hand by the database designer, or automatically derived from existing database schemas.
Translation
The translator builds a database whose schema implements the structure given by the extracted portions of the ontology. In addition, it generates view and constraint definitions which implement the semantics of concepts in the ontology with perfect fidelity and high efficiency. The user can guide the translator to omit some details for improved performance.
Java Object Oriented (OO)/Relational API
The database is exposed through a Java API. The API provides a simple object-oriented view of the ontology which will be familiar to all Java programmers. The API also provides a relation-based view for more sophisticated queries. Both enforce strong typing rules, which improves program correctness, makes programs easier to reuse, and speeds program development.
XSB-based Deductive Database
A deductive database (DDB) is about as close as databases are ever likely to get to ontologies, and translating from (part of) an ontology to a DDB requires, in general, the least loss of information. This is why it was decided to develop a translator for a DDB first, before a relational or object-oriented data model. The core of the deductive database is XSB, a main memory deductive database system developed at the State University of New York, Stony Brook. XSB itself lacks many features found in traditional database systems. To compensate for this, OW provides a database server built on XSB which provides transaction and recovery services, while taking advantage of the query processing efficiency of the DDB.
The system and method of the present invention include various improvements and variations over the system described by Peterson et al. In particular, the system according to the invention has both conceptual and implementation improvements over the Peterson et al. system including, but not limited to, those improvements described below.
Conceptual Improvements
Conceptual improvements were made to
Andersen William A.
Brinkley Paul M.
Engel Joshua F.
Peterson Brian J.
Hamilton Monplaisir
McGuireWoods LLP
Ontology Works, Inc.
Vu Kim
LandOfFree
Ontology for database design and application development does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Ontology for database design and application development, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Ontology for database design and application development will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3164543