=Paper=
{{Paper
|id=Vol-1979/paper-16
|storemode=property
|title=Semantic Consistency in Enterprise Models -- Through Seamless Modelling and Execution Support
|pdfUrl=https://ceur-ws.org/Vol-1979/paper-16.pdf
|volume=Vol-1979
|authors=Tomas Jonsson,Håkan Enquist
|dblpUrl=https://dblp.org/rec/conf/er/JonssonE17
}}
==Semantic Consistency in Enterprise Models -- Through Seamless Modelling and Execution Support==
https://ceur-ws.org/Vol-1979/paper-16.pdf
Semantic Consistency in Enterprise Models –
Through Seamless Modelling and Execution
Support
Tomas Jonsson1 and Håkan Enquist2
1
Genicore AB, S-411 09 Göteborg, Sweden,
tomas@genicore.se,
WWW home page: http://genicore.se
2
Department of Applied IT, University of Gothenburg, Sweden,
hakan.enquist@gu.se,
WWW home page: http://www.gu.se
Abstract. Semantic consistency, in context of a socio-technical view on
Enterprise Information Systems (EIS), concern consistency between En-
terprise Models, Conceptual Models expressed in professional language,
Information Systems models and the factual data manipulation and func-
tions implemented in runtime IT artefacts. CoreWeb, an environment for
modelling and testing conceptual models, implemented as an easy to use
web service, provide support for model editing and execution, semantic
consistency exploration and checking as well as producing documenta-
tion for either requirement specifications and/or for execution in model
driven runtime environments such as CorePro. CoreWeb modelling views
and features are presented in the paper and illustrated in the Demo com-
prising a walkthrough of a case supporting dialog and model verification
between users, domain experts, modellers and system designers.
Keywords: Information Systems, Phenomena Modelling, Semantic
Quality, Natural Language, Modelling Environment, Model Execution
1 Background
Conceptual models can be used to reflect relevant concepts of an enterprise and
play an important role in communication between users, domain experts, mod-
ellers and system designers, in the life cycle management of IT based Information
Systems (ITbIS).
If the Enterprise Information System (EIS) is regarded as a socio-technical
system i.e. a system of people and information artefacts, the semantic consistency
of a conceptual model as a foundation for ITbIS development, can only be fully
evaluated with the resulting ITbIS put in context of the socio-technical EIS. This
evaluation will of course only be relevant if the ITbIS implementation accurately
correspond to the semantic model.
Additionally, the socio-technical systems are comprised of various enterprise
actors with different semantic perspectives. If the semantic model is to be con-
Copyright © by the paper’s authors. Copying permitted only for private and academic
purposes.
In: C. Cabanillas, S. España, S. Farshidi (eds.):
Proceedings of the ER Forum 2017 and the ER 2017 Demo track,
Valencia, Spain, November 6th-9th, 2017,
published at http://ceur-ws.org
II
sistent with these perspectives, the semantic modelling facility should include a
perspective-modelling component.
The life cycle process, from conceptual model to target ITbIS, in traditional
approaches, involves several steps of manual transformations of what the concep-
tual model expresses. Each step involves several degrees of freedom for design
and implementation compromises resulting in deviations from the statements
of the conceptual model. However, with a semantic model driven ITbIS, i.e. se-
mantic model execution, free from manual transformation processes, consistency
between model and implementation could be guaranteed.
2 Approaches to conceptual model execution
Environments executing conceptual models in purpose of verifying conceptual
consistency exist, such as Merode [1] a modelling and model execution environ-
ment aimed at practitioners and students. Merode however does not support
multi perspectives modelling, such as suggested in [2].
Core Enterprise Architecture Framework (CoreEAF) is a process agnostic,
rule based, reference architecture and modelling method for multi perspectives
information systems, proven to be viable for large scale information systems
development [3] in combination with CorePro model and runtime environment
(a C++, SQL and XSLT based information systems platform).
3 CoreWEB
CoreWEB is an experimental (non commercial) modelling and execution WEB
service based on CoreEAF. CoreWEB is aimed at students and practitioners who
want to focus on enterprise phenomena modelling with multi-actor perspectives.
The CoreWEB project was initiated in 2015 with the key target to simplify
the ’model to system’ process as well as exploring various new modelling and
system quality aspects.
CorePros web based modelling component was partly reused and extended.
A new (and compared with CorePro) simplified system execution strategy was
designed, including Javascript code generation for phenomena model execution,
a model driven user interface engine with user interface model editing capability
and a web service platform.
CoreWEB, apart from being a WEB service, offers new features focusing
on semantic quality for both conceptual model and system. Adding professional
language sentences feedback (currently in Swedish and English). A semantic user
interface model ensures semantic consistency between model and user interface,
an auto layout system enforce a syntactically consistent user interface.
As an experiment and requirement elicitation driver, CoreWEB, with some
additions, has been used for a limited production system case (Lazarus, a live
action role play event association with 600 members).
Environment and demonstration videos are available from Association for
Model Enabled Information Systems. https://ameis.se/cml/index.htm
III
4 Demo of CoreWEB
The CoreWEB phenomenon type-modelling component has some different views
Fig. 1. The phenomenon map, is a two dimensional map of types of phenomena
and their basic relationships, used for model navigation. The definition of a
phenomenon type is comprised of attribute definitions. Types of phenomena
and their attributes, except Boolean attributes, should be labelled with nouns
or noun phrases, in stating both singular and plural forms.
In the phenomenon map, types of phenomena are displayed using the plural
noun form, since the model item represents a set of instances of a certain type.
Relation attribute labels are displayed according to their multiplicity, zero-one
or zero-many. From the model, a dictionary can be generated, defining each mod-
elled enterprise concept with an auto generated statement. Also when executing
the ITbIS, singular and plural noun forms are displayed appropriately, in Fig. 1
for example the 2 possessions list and the owner field.
Fig. 1. Views in CoreWEB environment. Top: Phenomenon map and attributes of
selected phenomenon. Middle: Labelling concepts and generated dictionary. Bottom:
Resulting runtime ITbIS
IV
4.1 CoreWEB Technical Architecture
The phenomena modelling component generates XML Fig. 2, which is trans-
formed by three XSLT processes into, natural language statements, executable
Javascript phenomena model equivalent and a user interface model, which is
kept consistent in respect to changes to the phenomena model. When system is
executed, the UI Engine will read the UI-model and interact with the executing
phenomena instance model according to the UI model.
Fig. 2. CoreWEB environment architecture
5 Notes and Comments
CoreWEB approach to enterprise modelling facilitates dialog, user involvement,
conceptual conflict resolution and rich expressions of enterprise knowledge as a
direct design basis for information management and ITbIS artefacts.
The resulting models, phenomenon and UI models, can be used as part of a
specification for ITbIS development regardless of implementation strategy or be
transferred to the CorePro production system development environment.
Future research includes enrichment of model semantics and natural language
model feedback by adding semantic variants of relations and phenomena and
experiments with data content and explanatory text and speech generation.
References
1. Sedrakyan G., Snoeck M. 2014: Lightweight Semantic Prototyper for Conceptual
Modeling, Advances in Conceptual Modeling: ER 2014, p298-302
2. Frank U. 2014: Multi-perspective enterprise modeling: foundational concepts,
prospects and future research challenges, Softw Syst Model (2014) 13:941-962
3. Jonsson T., Enquist H. 2015: CoreEAF – a Model Driven Approach to Information
Systems. Proceedings of CAISE Forum 2015, p137-144