Separating Conceptual and Visual Aspects in

Meta-Modelling

Simon Nikles and Simon Brander

FHNW – University of Applied Sciences Northwestern Switzerland

Riggenbachstr. 16, 4600 Olten, Switzerland

Abstract. ATHENE is a modelling environment which allows for creating

meta-models and models based on these, aimed at generating an ontology of the

modelled content. This work deals with the question on whether and how to

separate the conceptual part of a (meta-)model from its graphical representation.

We provide an overview on existing ontology- and meta-modelling approaches

compared to ATHENE and develop a conceptual basis for enhancing the meta

level of the tool.

1 Introduction

Nowadays, semantic technologies have matured and are now well established. There

usually are, however, two major drawbacks when working with such technologies:

First, the tools are often too complex for business users that do not think in terms of

concepts, instances and properties. Second, building ontologies is time-consuming

and not that easy: while ontology experts lack the insight into the business knowl-

edge, the domain expert lacks the expertise to create a formally correct description

(i.e. model) of her knowledge. To solve these issues, the University of Applied

Sciences Northwestern Switzerland is developing the ATHENE environment. The

ATHENE modelling environment applies a meta-modelling approach as described in

many sources, such as [7] or [4] to define domain-specific graphical notations and

generic operations to transform the end user's models into an ontology. The

metamodels themselves are represented as ontologies combining the conceptual

domain and the visualisation.

A weakness discovered when combining conceptual and visual aspects of a model

was found when similar individuals occur in more than one model or multiple times

in the same model: this results in multiple model-elements which actually mean the

same thing (e.g. with different positions). Further on, as each model-element has

individual visual attribute-values like position or size it cannot represent classes. By

attaching information on visualisation to the conceptual elements, obviously artificial

data (e.g. the position of a task inside a process model) is added to the ontology. This

data has, however, nothing to do with the real things that the ontology should

represent.

These reasons led to the intention to separate the visualisation from the actual

ontology model.

Nikles S. and Brander S. (2009).

Separating Conceptual and Visual Aspects in Meta-Modelling.

In Proceedings of the Joint Workshop on Advanced Technologies and Techniques for Enterprise Information Systems, pages 90-94

DOI: 10.5220/0002200800900094

 SciTePress

2 Related Work

The main differentiation of ATHENE compared to other approaches to model

ontologies is that it focuses on the modelling of domain specific notations and only

generates an ontology internally, whilst other tools like Dome

(DERI Ontology

Management Environment), Protégé

, or OntoEdit [13], [12] remain in the thinking of

concepts and properties and model them mainly as concept-trees. As graph-based

visualisation is mostly an additional option or the visualisation is predefined and

independent from the meaning of the type (e.g. the shape of the concept computer is

similar to the one of the concept person), the ontology (conceptual model) is clearly

separated from the notation.

Several tools which apply meta-modelling approaches are available. The main

distinction of ATHENE is its ontology orientation. Unlike other environments like

GME (Generic modelling environment) [9] or Eclipse GMF (Graphical Modelling

Framework) ATHENE does not aim at code generation. Whilst for example GMF has

a clear separation of notation and abstract syntax, several tools and approaches

directly attach the visual information (e.g. ADONIS [1], GME [9]).

The importance of separating the abstract syntax from the graphical notation is

pointed out by numerous authors. [2] mentions for example the reusability of

visualisation objects and advantage when changing notations. [10] clearly proposes to

map the notation onto the abstract syntax. In his E3-Model [5] shows how views and

presentation relate to models.

Different approaches do also exist for the actual separation. Examples are the

semantic visualisation approach [2], or the schema-based approach [3]. For ATHENE

a mapping-approach which in particular leans on the concept of the GMF will be

described in the next chapter

3 Mapping Approach in ATHENE

A distinct mapping of the visual elements to the abstract elements of a model enables

in particular reusability of visual elements, allows for different visual representations

of a specific element and offers an anchor to attach further definitions like

behavioural specifications or attribute mappings.

The resulting (partially simplified) meta

model of ATHENE is presented in (Fig. 1).

There may exist model-types that consist of object-types. Specialised object-types are

edges and container types which have predefined relational attributes. All elements

may have attributes. With this part of the meta

model, the abstract meta-models are

defined. Each model-type also refers to the mappings that connect the abstract

elements with certain visual elements, denoted as "element view" and specialised by

object-, relation- and container-view offering different visualisation properties. The

property mappings, which are attached to the element mappings offer possibilities to

influence the visualisation depending on values of an abstract element's attribute.

DOME: http://dome.sourceforge.net/

protégé: http://protege.stanford.edu/

Fig. 1. Conceptual meta

model with mapping.

Fig. 2. Simple metamodel based on the reworked meta

model.

Fig. 2 shows a very simple meta-model ("SimpleFlow") consisting of only one object-

type ("FlowObject") and an edge-type ("Arrow"). The abstract attribute 'Name' is used

as label of the visual elements through the property map ("MapLabel").

Finally Fig. 3 shows the model "Example" of type "SimpleFlow". It simply

contains of two flow objects (Element 1 and 2) and a relation between these objects.

Fig. 3. Example model based on the metamodel "SimpleFlow".

4 Conclusions

On the basis of application scenarios and related literature, several arguments were

worked out for the separation of abstract models and their visual representations.

These arguments especially concern flexibility and reusability and do confirm our

intentions. The only disadvantages found are concerned the simplicity and

redundancy, as in any approach, the separation obviously requires per conceptual

element at least another element for each of its visualisations and, for flexibility

reasons, even an additional mapping element.

In consequence, an extension of the existing meta

-model in ATHENE based on a

mapping approach was developed. The approach offers high flexibility and offers

potentials for further research. In the next step, a prototypical implementation will be

completed and tested.

The literature study also gave indications for further research topics. Namely, the

conceptual model could be revised on basis of OMG's Ontology Definition

Metamodel. Especially possible enhancements of variability and functionality, e.g.

more sophisticated approaches to define visual elements, dependency of properties

and reusing parts of the visualisation could be reached by approaches like the usage of

vector graphics and constraint languages as mentioned in [2] and [3]. Further on,

ATHENE does not support modelling rules at the moment, hence defining and

verifying constraints is a related important future task.

References

1. Fill, H.-G.: UML Statechart Diagrams on the ADONIS Metamodeling Platform. In:

Proceedings of the International Workshop on Graph-Based Tools (GraBaTs 2004),

Electronic Notes in Theoretical Computer Science, Vol. 127, No. 1, (2005) 27-36

2. Fill, H.-G.: Semantic Visualisation of Heterogenous Knowledge Sources. In: K.

Hinkelmann and U. Reimer, eds. Modellierung für Wissensmanagement – Workshop im

Rahmen der Modellierung 2006. Sonderdrucke der Fachhochschule Nordwestschweiz,

2006-W01, (2006) 17-27

3. Fondement, F., Baar, T.: Making Metamodels Aware of Concrete Syntax?. In: Model

Driven Architecture – Foundations and Applications, First European Conference, ECMDA-

FA 2005, Nuremberg, Germany, 7-10 November 2005. Springer-Verlag, Berlin, Heidelberg

(2005)

4. Geisler, R., Klar, M., Pons, C.: Dimensions and Dichotomy in Metamodeling. In:

Proceedings of the Third BCS-FACS Northern Formal Methods Workshop, September

1998. Springer-Verlag, New York (1998)

5. Greiffenberg, S.: Methodenentwicklung in Wirtschaft und Verwaltung. Verlag Dr. Kovač,

Hamburg (2004) 101ff

6. Junginger, S., Kühn, H., Strobl, R., Karagiannis, D.: Ein Geschäftsprozessmanagement

Werkzeug der nächsten Generation - ADONIS: Konzeption und Anwendungen.

Wirtschaftsinformatik, Vol. 42 No. 5, (2000) 392f

7. Karagiannis, D., Kuhn, D.: Metamodelling Platforms. In: K. Bauknecht and A. Min Tjoa

and G. Quirchmayer, eds. Proceedings of the Third International Conference EC-Web

2002, LNCS 2455. Springer-Verlag, Berlin, Heidelberg (2002) 182

8. Kosayba, B., Marvie, R., Geib, J.: Model Driven Production of Domain-Specific Modeling

Tools. In: 4th OOPSLA Workshop on Domain-Specific Modeling, Vancouver - Canada,

October 2004 (2004)

9. Ledeczi, A., Maroti, M., Bakay, A., Karsai, G., Garrett, J., Tomason, Ch., Nordstrom, G.,

Sprinkle, J., Volgyesi, P.: The Generic Modeling Environment [online] (2000). Available

from: http://www.isis.vanderbilt.edu/sites/default/files/GME2000Overview.pdf [Accessed

15 December 2008]

10. Object Management Group: OMG Unified Modeling Language (OMG UML),

Infrastructure, V2.1.2 (2007)

11. Object Management Group: Business Process Definition MetaModel, Volume II: Process

Definitions. Version 1.0 (2008)

12. Sure, Y., Erdmann, M., Angele, J., Staab, S., Studer, R., Wenke, D.: OntoEdit:

Collaborative Ontology Development for the SemanticWeb. In: Proceedings of the first

International Semantic Web Conference 2002 (ISWC 2002), 9-12 June 2002 Sardinia

(Italia). Springer-Verlag, Berlin, Heidelberg (2002)

13. Sure, Y., Studer, R.: On-To-Knowledge OntoEdit, EU-IST Project IST-1999-10132 On-To-

Knowledge, Deliverable 3, 2001 [online] (2001). Available from:

http://www.ontoknowledge.org/downl/del3.pdf [Accessed 5 January 2009]