1.3 Overview of the Report

The sections following this executive summary jointly constitute a comprehensive report on the

IDEF5 ontology capture method. Section 2 of the report provides a detailed discussion of the

conceptual foundations of IDEF5. It begins by tracing the roots of ontological inquiry, with

8

respect to engineering and manufacturing, to the classical philosophical tradition — also known

as “ontology” — of characterizing and classifying what ultimately exists. From this tradition

springs the central concepts of ontology: kinds (roughly, classes or types), properties, attributes,

relations, parts and wholes, and processes. Most of Section 2 is devoted to explicating and

relating these concepts. The section closes with a discussion of the need for a separate ontology

method distinct from existing methods. (The relation of the ontology capture method with

existing IDEF methods is discussed in Subsection 1.3.)

Section 3 provides a practical method for the construction of ontologies. Ontology development

requires extensive iterations, discussions, reviews, and introspection. Knowledge extraction is

usually a discovery process and requires considerable introspection. It requires a process that

incorporates both significant expert involvement as well as the dynamics of a group effort.

Given the open-ended nature of ontological analyses, it is not prudent to adopt a “cookbook”

approach to ontology development. We recommend the use of a general procedure along with a

set of useful guidelines. Section 3 describes the mechanics of such a process for potential IDEF5

ontology developers.

In brief, the IDEF5 ontology development process consists of the following five activities.

1. Organizing and Scoping This activity involves establishing the purpose, viewpoint,

and context for the ontology development project and assigning roles to the team

members.

2. Data Collection This activity involves acquiring the raw data needed for ontology

development.

3. Data Analysis This activity involves analyzing the data to facilitate ontology extraction.

4. Initial Ontology Development This activity involves developing a preliminary

ontology from the acquired data.

5. Ontology Refinement and Validation This activity involves refining and validating the

ontology to complete the development process.

Although the above activities are listed sequentially, there is a significant amount of overlap and

iteration between the activities. These activities, and their interconnections, are described in

detail in Section 3.

9

Section 4 contains a description of the IDEF5 ontology languages. There are two such

languages: the IDEF5 schematic language and the IDEF5 elaboration language. The schematic

language is a graphical language that has been specially tailored to enable domain experts to

express the most common forms of ontological information, especially with the aid of an

automated ontology capture tool. This enables average users both to input the basic information

needed for a first-cut ontology and to augment or revise existing ontologies with new

information.

There is a price for the relative ease of use of the schematic language, viz., that it lacks the full

expressive power needed to capture general ontology information. To capture such information

is the purpose of the IDEF elaboration language. The elaboration language is a structured text

language with the full expressive power of first-order logic and set theory. This enables a user to

express virtually any condition, or relation, or fact that one might need to express to characterize

a given kind of thing, or property, or relation, or process found in a domain. In addition to set

theoretic constructs, the language also includes specialized constructs for expressing ontology

information in the particular format of IDEF5. This makes for easy translation from the

schematic language into the elaboration language, and vice versa, insofar as that is possible.

Finally, the report concludes with two appendices. The first is the (current) IDEF5 relation

library of reusable ontology elements. This library is a rich repository of information consisting

of a set of characterizations of (i.e., definitions and axioms for) commonly used relations. It

provides a repository of formally defined and characterized relations that can be reused and

customized in a particular project. The relation library itself is a specialized ontology: an

ontology of commonly used relations. The motivation for this library grew out of the previously

mentioned analogy with software engineering. Often in software development, the same kinds of

routines are used again and again in different programs by (in general) different programmers.

The development of ontologies will face the same sort of problem. It is likely that the same or

similar relations will appear in a number of different ontologies. The role of a library of relations

such as the one presented in Appendix A will be to enable modelers to reuse and customize

relations that have been defined in previously captured ontologies. The library can also be used

as a reference for the different ways to define and characterize relations and illustrative examples

of the use of the IDEF5 elaboration language. All definitions and characterizing axioms in the

library have been written using the IDEF5 elaboration language. Thus, the library can also serve

as a useful learning tool for mastering the IDEF5 elaboration language. Finally, the library is

extensible in that any relation that has been formally defined and characterized may be added to

it.

10

The second appendix consists of the BNF specification of the IDEF5 elaboration language to

ensure that the language is well-defined. A glossary for the report follows this appendix.