Conceptual modeling languages such as Entity-Relation-Models (ERM), Unified Modeling Language (UML), and Object-Role-Modeling (ORM) are used especially for modeling business domains that then are mapped mainly to relational databases. Do you use any similar tools when creating an ontology or directly start creating classes and properties, for instance in Protegé? I found some research papers on conceptual modeling for OWL, but I doubt that these prototypes are much used in practice, although academic tools like DOGMA look interesting. To answer my own question: my most important tools are pencil and paper for mind-mapping, and I sometimes use ORM before writing down OWL.

P.S: The term "conceptual modeling" seems to have different meaning in database community and semantic web community. In doubt, you may remove the "conceptual" and think about any explicit tools to create artifacts that are translated to RDF ontologies in a next step.

P.P.S: Some candidates for conceptual modeling in ontology engineering, mentioned so far:

  • Mindmaps
  • Conceptual Graphs
  • TopBraid Composer
  • NORMA (fact based / object role modeling)
  • Directly use OWL (possibly with support of a specific editor)

There is a number of tools from academia, but in practice they seem to have little relevance.

asked 24 Aug '11, 07:54

Jakob's gravatar image

Jakob
1.9k211
accept rate: 10%

edited 26 Aug '11, 03:39

3

I consider ontology languages, such as OWL, as conceptual modeling languages, just as ERM or UML are. I don't see a need for a "higher level" language for writing down what I want to express before a translation into OWL. Maybe OWL as it is today misses some features to express certain things, but then I would rather ask for an extension of OWL (same conceptual level, but richer) instead of looking for another "higher-level" language. In particular, I do not consider UML an appropriate modeling language on top of OWL, because it doesn't directly support the logic-based way OWL works.

(25 Aug '11, 02:24) Michael Schn... ♦ Michael%20Schneider's gravatar image

A crucial property of conceptual modeling languages is they have a graphical notation. From a theoretical standpoint it may make no difference whether you write OWL in RDF/XML, Turtle, Manchester syntax, or carved in stone. But conceptual modeling depends on usability, which is rather poor if you directly use OWL. However it is common practice not to use any explicit conceptual modeling language at all, which applies if you directly use OWL from the beginning.

(25 Aug '11, 03:16) Jakob Jakob's gravatar image

Gruber defined ontology in information science as:

an ontology is an explicit specification of a conceptualization

(see Thomas R. Gruber. Toward Principles for the Design of Ontologies Used for Knowledge Sharing. Toward Principles for the Design of Ontologies Used for Knowledge Sharing, 43:907–928, 1993)

So one can view a specific ontology as a conceptual model. Pencil and paper are really good tools for mind mapping in this context. The important difference one has always to keep in mind, that the other conceptual model types you mentioned are closed views and ontologies are open ones. Regarding specific tools: in TopBraid Composer (Maestro Edition) you have a graphical editor for design your conceptual model, which is really good from my POV.

PS (EDIT): you may have a look at some of John F. Sowa's writings, e.g.,

EDIT2: "Conceptual Graphs and the Semantic Web " is a good comparison of both concepts made by Tim Berners-Lee.

EDIT3: "Towards an Ontology for Conceptual Modeling" (McCusker et al., 2011)

link

answered 24 Aug '11, 10:41

zazi's gravatar image

zazi
3.4k213
accept rate: 13%

edited 28 Aug '11, 11:03

Gruber's definition is far too broad, my question aimed at RDF ontologies, which are a small subset of "explicit specifications of a conceptualization". Thanks for highlighting the open world/closed world issue. TopBraid looks promising, but it seems to only support UML-like Class diagrams. The visualization of mappings and scripts is a good idea.

(25 Aug '11, 03:31) Jakob Jakob's gravatar image

Of course, ontology diagrams are looking quite similar to UML class diagrams. However, this is due their nature in representing conceptual models. A common mistake is (from my POV) that ontology diagrams are occasionally designed as UML class diagrams and people are thinking with the restrictions of UML class modelling when designing ontologies in that way.

(25 Aug '11, 04:36) zazi zazi's gravatar image

Sowa's Conceptual Graphs seem like a good candidate for conceptual modelling also for ontology engineering. It's only the question whether they are actually used in practice.

(26 Aug '11, 03:26) Jakob Jakob's gravatar image

@Jakob: then you may have a look at the link of the Conceptual Graph/Semantic Web comparision I added to my answer ;)

(26 Aug '11, 04:28) zazi zazi's gravatar image

The DOGMA method and tools have been commercialized by the Vrije Universiteit Brussels’ spin-off company Collibra (http://www.collibra.com/ Business Semantics Management, Studio and Glossary), you can always try contacting them for a trial for research purposes.

DOGMA is a fact-oriented approach for modeling ontologies based on Object Role Modeling (ORM, which NORMA supports). In DOGMA, a community comes to a shared agreement by communicating facts in natural language. Those facts are binary, meaning you can read it in two directions. Both roles must be made explicit. This is different from other formalisms where you can model concepts and properties separately (and even not necessarily need to declare domains and ranges of properties). Note that taxonomy trees can be constructed with the relation “is a / subsumes”.

After a consensus is achieved, the ontology (in DOGMA) can be transformed/implemented in other formalisms such as OWL. Business Semantics Studio does this via OMG’s MOF framework, others (such as Jarrar) have publishes papers on how to translate ORM (and therefore also DOGMA ontologies) into DL.

In short, DOGMA does not aim to replace formalisms more suitable for reasoning or processing, instead it’s a mean to model an ontology using facts and natural language which can be translated into other formalisms later on.

link

answered 29 Aug '11, 11:23

chrdebru's gravatar image

chrdebru
111
accept rate: 0%

Your answer
toggle preview

Follow this question

By Email:

Once you sign in you will be able to subscribe for any updates here

By RSS:

Answers

Answers and Comments

Markdown Basics

  • *italic* or _italic_
  • **bold** or __bold__
  • link:[text](http://url.com/ "Title")
  • image?![alt text](/path/img.jpg "Title")
  • numbered list: 1. Foo 2. Bar
  • to add a line break simply add two spaces to where you would like the new line to be.
  • basic HTML tags are also supported

Tags:

×79
×14
×8

Asked: 24 Aug '11, 07:54

Seen: 2,274 times

Last updated: 29 Aug '11, 11:23