Authors:
Chris Partridge
1
;
Sergio de Cesare
2
;
Andrew Mitchell
3
;
Ana León
4
;
Frederik Gailly
5
and
Mesbah Khan
6
Affiliations:
1
University of Westminster and BORO Solutions Ltd., United Kingdom
;
2
University of Westminster, United Kingdom
;
3
BORO Solutions Ltd., United Kingdom
;
4
Universitat Politècnica de València, Spain
;
5
Ghent University, Belgium
;
6
Tullow Oil plc., United Kingdom
Keyword(s):
Enterprise Modelling, Computation Independent Model, Ontology, Agentology, Essential Indexical, De Se, De Re, De Se Knowledge, De Re Knowledge, Deitic Centre, Agent-Neutral, Agent-Relative.
Related
Ontology
Subjects/Areas/Topics:
Frameworks for Model-Driven Development
;
Languages, Tools and Architectures
;
MetaModeling
;
Methodologies, Processes and Platforms
;
Model-Driven Architecture
;
Model-Driven Software Development
;
Models
;
Paradigm Trends
;
Software Engineering
Abstract:
Data integration of enterprise systems typically involves combining heterogeneous data residing in different sources into a unified, homogeneous whole. This heterogeneity takes many forms and there are all sorts of significant practical and theoretical challenges to managing this, particularly at the semantic level. In this paper, we consider a type of semantic heterogeneity that is common in Model Driven Architecture (MDA) Computation Independent Models (CIM); one that arises due to the data’s dependence upon the system it resides in. There seems to be no relevant work on this topic in Conceptual Modelling, so we draw upon research done in philosophy and linguistics on formalizing pure indexicals – ‘I’, ‘here’ and ‘now’ – also known as de se (Latin ‘of oneself’) or the deitic centre. This reveals firstly that the core dependency is essential when the system is agentive and the rest of the dependency can be designed away. In the context of MDA, this suggests a natural architectural l
ayering; where a new concern ‘system dependence’ is introduced and used to divide the CIM model into two parts; a system independent ontology model and a system dependent agentology model. We also show how this dependence complicates the integration process – but, interestingly, not reuse in the same context. We explain how this complication usually provides good pragmatic reasons for maximizing the ontology content in an ‘Ontology First’, or ‘Ontology then Agentology’ approach.
(More)