Theoretical Foundations and Engineering Tools for Building Ontologies as Reference Conceptual Models

Paper Title: 
Theoretical Foundations and Engineering Tools for Building Ontologies as Reference Conceptual Models
Giancarlo Guizzardi
Perhaps the most fundamental notion underlying the desiderata for a successful Semantic Web is Semantic Interoperability. In this context, ontologies have been more and more recognized as one of the enabling technologies. This paper defends the view that an approach which neglects the role of ontologies as reference conceptual models cannot meet the requirements for full semantic interoperability. The paper starts by offering an engineering view on ontology engineering, discussing the relation between ontologies as conceptual models and ontologies as codification artifacts. Furthermore, it discusses the importance of foundational theories and principles to the design of ontology (conceptual) modeling languages and models, emphasizing the fundamental role played by true ontological notions in this process. Finally, it elaborates on the need for proper tools to handle the complexity of ontology engineering in industrial scenarios and complex domains. These tools include ontological design patterns as well as well-founded computational environments to support ontology creation, verification and validation (via model simulation).
Full PDF Version: 
Submission type: 
Responsible editor: 

Review 1 by Oscar Corcho:

This paper presents an analysis of how the Ontological Engineering field should build upon some of the work that has been done in other areas such as Software Engineering, especially in what respects to the separation of concerns among the different types of models that can be managed throughout the ontology development lifecycle.

In my opinion, the paper is well structured and well suited for the inaugural issue of this journal, showing some aspects that may generate controversy and discussion in the community with respect to how to approach the field of Ontological Engineering in the future.

My only consideration is that I think that it would be useful as well to reflect on how ontology engineering methodologies address this separation of concerns, and how the tools that have been developed to give support to them could be adapted, if needed. Besides, a comparison with respect to approaches followed by many ontologists, such as the use of competency questions as a guiding principle for the development and testing of ontologies, would be nice.

In terms of the transformations needed to move from one model in a level to another one in a different level, the paper does not get into depth on the problems that can be found in this respect, taking into account the different KR formalisms in which ontology languages can be based on, or the differences in expressivity that may exist among them and between them and the upper-layer models.

Review 2 by Pascal Hitzler:

I read this position paper with much pleasure. It provides a refreshing perspective while it constructively questions some aspects of ontology engineering which have happened to become established in the community. It makes an excellent contribution to the special issue.

Of particular importance is, in my opinion, that "Ontologies as Reference Conceptual Models" has been singled out as a rather distinct use case from ontologies for automated reasoning, and one of the main points put forth in the paper, namely that the first-mentioned use of ontologies requires distinct approaches to ontology engineering, makes a lot of sense to me.

There is one aspect, though, where I would be interested in understanding how it fits into the perspective of the author: What is the role of (formal) semantics? In particular, the author argues that reference ontologies should be modeled in expressive (high-level) languages, and could then be instantiated in several less expressive ontology languages (such as OWL or F-Logic). If we think about the latter, though, then these languages cannot be separated from their specified (model-theoretic) semantics. This, then, seems to destroy the desired "interoperability" of the ontology, since it could have realizations under different (and perhaps contradictory) semantics. I'd like to hear the author's perspective on this.

some minor corrections:

page 2 left: need an ontology conceptual modeling language that assist*s* modelers

page 2 right bottom: OWL Lite has not been carried over to version 2 of the OWL standard, so it could be replaced here (e.g. by mention of OWL Full). RIF could also be mentioned.

page 3 left line -10: received much attention *in* the context

page 3 right line 3+4: Finally, there is *an* important additional aspect which I would like to *draw* attention *to*

page 3 right line -15: to be able *to* produce

page 4 left line 1: e.g., *p*etroleum (twice)

line 3: this -> these

half-way down: but they *should* rather reflect

bottom: (iii) -> (iv)

page 4 right half-way down: intricateness -> intricacies

page 5 left top: I predict that *they* will hardly

half-way down: primitives such a*s* classes

right, two-thirds down:
* Regarding (i), OntoUML incorporates
* embeds constraints such *as*:
* [9] reports on a*n* implementation of *an* OntoUML graphical editor

page 6 left top:
* why this is the case is ontological
* as most language speakers do

* transformed in*to* a specification
* creation of *the* model simulator

In the rest of page 6, there are several occurrences of "state of affairs" which should probably be "states of affair" - later it is written as "states-of-affairs".

page 6 right one-third down:
* supertype _> subtype (?)
* violate any of these rule*s*

two-thirds down:
* into consider*ation*
* to the OntoUML editor *was presented*

page 7 left, half-way down: called attention for -> brought to attention (?)

right, top: and sensitivity to *the* need of