From Sustainable Information System with a Farandole of Models to Services - Intentional Perspectives on Information Systems Engineering - page 10

governance tasks consists to assure that these models are coherently articulated and
to manage the IS development to achieve this result.
Furthermore, the IS governance has to face multiple factors provoking transfor-
mations of one or several models of the farandole. This situation can be overcome
if the meta-models support the evolution of their models. A meta-model supports
evolution if it includes a complete set of mechanisms to manage the evolution of its
models. That means that a meta-model is described as a database schema with its
integrity rules to assure the coherence of the facts to be stored and which constitute
a model. The mechanism of evolution executes an atomic modification such as cre-
ating, deleting or updating of a stored fact and verifies all the involved integrity
rules of the meta-model. The set of mechanism is complete if all the primitive
actions on the database - creating, deleting, updating - are taken into account [ 2] .
Consequently, the systems underlying IS must include a complete set of mechanisms
of evolution, like [3] .
After or during the evolution process of a model, the IS evolution process requires
to analyse the consequences of its evolution on the other models, which are articu-
lated with it. Indeed the general question of the IS evolution concerns not only the
evolution of one model but also the evolution of the farandole of the models.
At the commencement of the IS domain, the pioneers tried to create a whole
methodology to develop an information system (e.g. Merise [ 33] ). Their results still
provide a framework for IS development. Besides, some very substantial works
were finalised by developing computer assisted information system engineering
environments (e.g. Rubis [30] ).
Nevertheless these methodologies became too imprecise to cover the whole life
cycle of an IS development. That is due to the huge variety of situations that design-
ers have to face to. A promising way is to use as a pivot the concept of situation
(e.g. [ 19, 36] ) for the investigation of which type of method, or rather an assembly
of method components, is the most appropriate for a given IS development. Several
approaches (e.g. [ 24] introduce the notion of method component (also fragment or
chunk) and propose different ways to assemble them into a situation-specific method
for IS development. Besides, there are also attempts to implement an information
system to manage the activities around an IS [ 17] .
3.6 Conclusions on IS Worlds
The IS world is in fact composed of several worlds: activities, conceptual, infor-
matics, ontology, governance. These worlds have their own properties and their
transformations are independent on each other. For a given IS, it is important to
assure the entente between all the models obtained in all these worlds. That is the
key role of the conceptual world to assure that all these models constitute an artic-
ulated and well-formed farandole. Moreover, this entente has to be dynamic: the
evolution of one model must be coherent with the other models, which perhaps,
also have to be modified in order to keep the articulated farandole consistent. The
IS is then sustainable .