semantic links between concepts. More precisely,
the concept Train is a composition of wheels,
engine, and an aggregation of wagons. These
semantic links have strong impacts on optimising
data. Indeed in our previous example the
composition implies that there cannot be instances of
the concept Engine without instances of the concept
Train. Deleting an instance of concept Train implies
that all dependent instances (Engine) will be
removed. So, the deletion process is optimised. In
the other hand, the aggregated instances (Wagon)
will not be deleted in the case of aggregation
between concepts.
Generalisation and specialisation relations are
used to factorise data. In the generalisation case,
common attributes are gathered in a general concept.
For example, attributes such as firstname and
lastname are common to the concepts driver and
mecanician. These two attributes are migrated to the
concept Person to factorise data, avoiding
duplication of their definition in the concepts driver
and mecanician.
4 CONCLUSIONS
We saw in this paper that we can facilitate
communication between people using a common
formalism, in our case using UML as an abstraction
of any technical language. UML is a powerful and
flexible modeling language and XML became a
standard for data interchange on the Web. The use of
these two technologies addresses interoperability
and a standard means of exchanging and defining
models. Therefore, we have shown how UML
profiles facilitate the definition of an XML Schema
model and fill the lack of semantic expressiveness.
Our follow-up work will consist of the
enrichment of MDM-p and the creation of a
complete data modeling language for the Master
Data Management module in the following two
ways: (i) a graphical language using UML modeling,
(ii) a specialised language using UML meta-
modeling. This language will take into account the
definition of constraints and the validation of models
using OCL (Object Constraint Language) (OMG,
2002) (Bazex, 2003).
REFERENCES
Abiteboul S., Cluet S., Ferran G., Rousset M.-C., 2002.
The Xyleme Project. Computer Networks 39.
Apvrille L., De Saqui-Sannes P., Khendek F., 2006.
TURTLE-P: A UML Profile for the Formal Validation
of critical and Distributed Systems. SoSym (Software
and System Modeling) Journal, Springer, ISSN 1619-
1366, pp. 1-18, July.
Bazex P., Bodeveix J-P., Millan T ., Le Camus C.,
Percebois C., 2003. Vérification de modèles UML
fondée sur OCL. INFORSID. Nancy, France. pp. 185-
200.
Garcia-Molina H., Papakonstantinou Y., Quass D.,
Rajaraman A., Sagiv Y., Ullman J., Widom J., 1995.
The STIMMIS approach to mediation: Data Models
and Languages. NGITS (Next Generation Information
Technologies and Systems), Naharia, Isreal, June 27-
29.
Greenfield J., 2001. UML Profile For EJB. Rational
Software Corp, May.
Lamolle M., and Zerdazi A., 2005. Intégration de Bases de
données hétérogènes par une modélisation
conceptuelle XML, In Conférence sur l’Optimisation
et les Systèmes d’Information (COSI’05), pp.216-227.
Ober I., Graf S., 2005. Timed annotations in UML
accepted to STTT, Int. Journal on Software Tools for
Technology Transfer Springer Verl.
OMG/MOF, 2000. Meta Object Facility (MOF)
Specification, OMG Document formal/2000-04-03,
http://www.omg.org/technology/documents/formal/mo
f.htm.
OMG, 2002. “CORBA specifications”.
http://www.omg.org/technology/documents/formal/pr
ofile_corba.htm.
OMG, 2002. Response to the UML 2.0 OCL.
http://www.omg.org/docs/ad/02-05-09.pdf
Orchestra Networks, 2000.
http://www.orchestranetworks.com
Pilone D., Pitman N., 2006. UML 2.0 in a Nutshell,
O’Reilly.
ICEIS 2008 - International Conference on Enterprise Information Systems
464