Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Theory and applications of ontology : computer applications
Poli R., Healy M., Kameas A., Springer Publishing Company, Incorporated, New York, NY, 2014. 576 pp. Type: Book (978-9-400793-66-8)
Date Reviewed: Oct 22 2015

“The ultimate goal of ontology [as a branch of philosophy] is to provide a definitive and complete classification of entities and the relations between those entities in all spheres of reality” (John Davies, p. 197 of this book). In information technology, we are interested in specific spheres of reality, sometimes also known as “universes of discourse” or “application domains.” This is what (business) modelers do, even those who do not explicitly use the term “ontology.” Thus, many modelers and model users may feel that they are in a familiar territory when they encounter the 24 chapters of this book.

Starting with the enjoyable preface, we see that “despite their different languages and different points of departure ... ontologies in knowledge engineering and ontologies in philosophy ... have numerous problems in common and ... seek to answer similar questions. ... Engineers and philosophers must devise ways to talk to each other.” Examples of such successful communications are relatively well known, notably, the use of Bunge’s ontology [1] in modeling [2,3 and elsewhere], and an international standard, the Reference Model of Open Distributed Processing (RM-ODP) [4], which is explicitly about “the expression of what exists, where it is, and what it does.” More generally, these patterns of thinking are not new: for example, Y. Manin, an outstanding 20th (and 21st) century mathematician, observes that in the culture of definitions developed by mathematicians, “many efforts are invested into clarification of … semantics of basic abstract notions and … of their interrelations, whereas the choice of words … and notations for these notions is a secondary matter” [5]. Regretfully, the only chapter where Bunge’s ontology is discussed is chapter 20 (on ubiquitous computing) by Goumopouls and Kameas, but even there relationship semantics is not emphasized; RM-ODP is not mentioned anywhere in the book, although already in the introduction we may be pleased to read that ontology is not the same as a data dictionary and that purely syntactic media of keywords and icons are not sufficient.

The authors distinguish among upper ontologies, mid-level ontologies, and specific domain ontologies. The first two are sometimes called “foundational ontologies.” The number of existing ontologies is huge: Loebe in chapter 3 quotes a 2007 paper stating that in the context of the semantic web effort, there exist “tens of thousands of OWL ontologies.” However, a language-oriented approach to ontologies is grossly inadequate (as noted, for example, by Loebe with respect to OWL); Poli and Obrst, the authors of chapter 1, properly observe that users’ imagination typically “is constrained by their current systems.” Regretfully, Poli and Obrst also claim that ontology from the perspectives of computer science originated approximately in 1991, which is clearly not the case because, for example, Wand and Weber used Bunge’s ontology already in 1988 and did not constrain their ontology at all: for an important example, the concept of emergent properties substantially used in [1-3] (and in [4]) is still alien to most current systems (and not mentioned in most chapters of this book).

Many chapters of the book survey various perspectives of ontology theory, architecture, constructs, and applications. A few chapters spell out such notions as clear conceptualizations of domains or of specializing/generalizing/aggregating existing concepts in developing such conceptualizations – this is what business modelers do in their everyday activities. For example, Kotis and Vouros in chapter 7 properly criticize machine-oriented conceptualizations that “cannot be further manipulated or even (in some cases) be inspected by domain experts” and reiterate the following most important issues in ontology engineering: allow an eclectic way (that is, any approach/combination of approaches) for ontology development, use abstraction by emphasizing the natural way to interact with conceptualizations, support means for conversation/collaboration/criticism in using/exchanging/evaluating ontologies, and support synonyms/homonyms in uncovering human-intended semantics. Such approaches are obviously preferable to the emphasis on machine processing where, in particular, “all” entities are “equally important.” Therefore, it is necessary to support a rigorous approach to information semantics, and especially to relationship semantics, not based on “meaningful names.” Proper handling of synonyms/homonyms, while essential and specifically mentioned in several chapters including chapter 19 on ontologies for e-government, is only the first step in the right direction. As observed by Davies (chapter 9 on “Lightweight Ontologies”), “lack of an explicit semantic model can lead to different (implicit) semantics being ascribed by different programs,” and of course by different human users. Good business modelers are well aware of this.

Several authors emphasize the need for a modeling language to support the system of concepts of a domain ontology and note that for popular modeling languages (and tools), this is often not the case. This observation is not new: for example, the semantics of several important concepts described in [1-3] or in such an international standard as [4] is also not supported by popular modeling languages.

Chapters 8 and 10 to 14 discuss various foundational ontologies. Of note, chapter 10 by Fellbaum clearly presents WordNet and does not require any specific prerequisites. Pease and Li in chapter 11 claim that “it has long been a goal in computer science for users to communicate with computers in human language,” but they use a very restricted subset and acknowledge that full-text understanding is “unlikely to arrive soon.” Indeed, as Dijkstra observed awhile ago, “so-called ‘natural language’ is wonderful for the purposes it was created for, such as to be rude in, to tell jokes in, to cheat or to make love in (and Theorists of Literary Criticism can even be context-free in it), but it is hopelessly inadequate when we have to deal unambiguously with situations of great intricacy ... which unavoidably arise in such activities as legislation, arbitration, mathematics or programming” [6]. Borgo and Masolo in chapter 13 tell us that “there cannot be a unique standard or universal ontology for knowledge representation” and properly promote a library of foundational ontologies reflecting different ontological choices.

Chapters 15 to 20 are on domain-specific ontologies. Notably, Kelso, Hoehndorf, and Prüfer in chapter 15 refer to serious and justified criticism of the Gene Ontology, specifically due to its reliance on tacit assumptions in general and on such “meaningful names” as “part-of” in particular (“the definition of the ‘part-of’ relationships was only added after a major part of the Gene Ontology had already been developed”) They also emphasize the importance of properly specified relationships including “part-of,” and note that axioms for the “part-of” relationship stated for the Open Biomedical Ontologies relationship ontology (reflexivity, transitivity, and anti-symmetry) are semantically weak (again, emergent properties of the composite [1-4] are not discussed). The authors also refer to problems due to the proliferation of synonyms and homonyms, and due to definitions of fundamental biological concepts (such as “gene”) that are not universally agreed upon, and properly distinguish between ontologies for idealized and “real-life” domains. Somewhat similarly, Bateman in chapter 17 notes the need to “augment the … lexical information with stronger semantics,” and “to avoid the well-known knowledge representation problem of ‘is-a’ overloading,” and even that “the categories involved ... have not yet been sufficiently understood.”

It would be interesting and instructive to create a business model of business ontologies, starting with the observations by Rittgen in chapter 18, with clearly specified relationship semantics (regretfully, in Fig. 18-3, all relationships are described using only the name “refers-to”). Rittgen also mentions Bunge’s ontological framework as one of the roots of enterprise ontology, but does not go into any details.

I would disagree with the claim in chapter 19 that “business people are not familiar with relating concepts.” Contrariwise, business experts can be familiarized with semantics of generic relationships, such as composition and subtyping, very quickly and deal with these relationships perfectly well (see, for example, [7]).

The last four chapters of the book are about category theory (CT) “as a mathematics for formalizing ontologies.” While CT is motivated and explained to a certain extent, the CT chapters, with the exception of chapter 24 by Johnson and Rosebrugh (who emphasize concepts more than technical details) and some examples in chapter 21 by Healy, require quite serious mathematical prerequisites. I prefer the “more applied” presentation for uninitiated readers in [8], where CT concepts are introduced only as needed and where a huge amount of examples from biological, social, and other systems is used to illustrate the narrative. Of note, Johnson and Rosebrugh observe that the requirement to check for semantic equivalence is “unfortunate but unavoidable” and that “whether ... mappings are meaningful is a semantic question that requires domain knowledge”; this is often ignored but (should be) well known to business modelers. Those of us who stress that objects do not exist in isolation will read with pleasure that “fundamentally the information about an instance of a class is contained in the way that that instance is related to other instances of other classes” (Johnson and Rosebrugh).

I would like to sum up with the final sentence of the final chapter 24 by Johnson and Rosebrugh: “It’s always pleasing when theoretical developments yield practical advantages and new insights as well as stronger theory.”

Unfortunately, the book has no index.

Reviewer:  H. I. Kilov Review #: CR143877 (1601-0033)
1) Bunge, M. Treatise on basic philosophy. Vol. 3: ontology. Reidel, Dordrecht, Netherlands, 1977.
2) Wand, Y. A proposal for a formal model of objects. In Object-oriented concepts, languages, applications, and databases. Edited by Kim, W.; Lochovsky, F. H. ACM Press, New York, NY, 1989, 537-559.
3) Wand, Y.; Weber, R. An ontological evaluation of systems analysis and design methods. In Information systems concepts: an in-depth analysis. Edited by Falkenberg, E. D.; Lindgreen, P. Elsevier, Amsterdam, Netherlands, 1989, 79-107.
4) ISO/IEC. Open Distributed Processing - Reference Model: Part 2: Foundations (ITU-T Recommendation X.902 ISO/IEC 10746-2), 1995.
5) Manin, Y. Mathematics as metaphor. American Mathematical Society, Providence, RI, 2007.
6) Dijkstra, E.W. Foreword https://www.cs.utexas.edu/~EWD/transcriptions/EWD12xx/EWD1238.html (10/16/2015).
7) Garrison, J. Business specifications: using UML to specify the trading of foreign exchange options. In Proc. of the 10th OOPSLA Workshop on Behavioral Semantics (Back to Basics). Baclawski, K. and Kilov, H., Eds. Northeastern University, Boston, MA, 2001, 79–84.
8) Ehresmann, A.; Vanbremeersch, J. P. Memory evolutive systems: hierarchy, emergence, cognition. Elsevier, New York, NY, 2007.
Bookmark and Share
  Reviewer Selected
Featured Reviewer
 
 
Ontologies (I.2.4 ... )
 
 
Knowledge Representation Formalisms And Methods (I.2.4 )
 
Would you recommend this review?
yes
no
Other reviews under "Ontologies": Date
Cohesion and coupling metrics for ontology modules
Oh S., Yeom H., Ahn J. Information Technology and Management 12(2): 81-96, 2011. Type: Article
Dec 16 2011
Ontology-based application integration
Paulheim H., Springer Publishing Company, Incorporated, New York, NY, 2011.  286, Type: Book (978-1-461414-29-2)
May 24 2012
A study on development of cognitive support features in recent ontology visualization tools
Ramakrishnan S., Vijayan A. Artificial Intelligence Review 41(4): 595-623, 2014. Type: Article
Jun 9 2014
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy