Method Engineering: A Service-Oriented Approach - Intentional Perspectives on Information Systems Engineering - page 333

Principle No 2: Variability and contextual knowledge. Two ways are available
to deal with goal satisfaction:
1. Propose one generic solution which can be applied in any case but which is not
the most efficient one in all the situations.
2. Propose several solutions, each one being relevant in a specific case.
We choose to adopt the second way: it will allow services to propose the
most relevant solution in each case. Then, a given goal can be achieved in dif-
ferent ways. Each way is more or less suitable according to consumers' context.
Contextual knowledge is very useful for discriminating the different manners to sat-
isfy the goal. We call “variability” the ability to propose several solutions for one
goal.
From the provider point of view, service modeling requires variability specifi-
cation mechanisms and contextual knowledge capture. From the consumer point of
view, a goal expression must be completed with contextual constraints which impact
the choice of the solution satisfying the goal.
Principle No 3: Semantic description of services. One of the difficulties for
services' consumers when discovering and invoking services is to express their
requirements with a specific unnatural language. This induces a gap between
consumers' knowledge and services specifications. In order to reduce this gap,
we propose to specify services with semantic data and to use “shared” ontolo-
gies. Service specification requires knowledge on goals, processes, actors and
objects.
From the provider perspective, ontologies should be used for specification of
semantic data that describe services. From the consumer perspective, ontologies
should be used to express requirements when discovering or invoking services.
Principle No 4: Dynamic adaptation. We consider that a large variety of
consumers can access and use a service, so service delivery should support per-
sonalization. Thus, services should take into account data about consumers that use
them. Usually, those data are expressed in users' profiles. By nature, users' profiles
evolve through time and, then, must be considered dynamically.
For the provider point of view, specification of services should weave know-ledge
about skills required to correctly use them. For the consumer point of view, a profile
should be updated through time for a relevant dynamic personalization.
Principle No 5: Dynamic composition. Some goals being complex, solutions
provided by a service may require satisfaction of sub-goals. In this case, a service
can delegate satisfaction of these sub-goals to other services.
From the provider perspective, specification of some solutions provided by ser-
vices must refer to goals which are satisfied by other services. From the consumer
perspective, this aspect should be as “transparent” as possible.
Due to goal orientation, variability and contextualization, the dynamic composi-
tion can be automatically computed by a service without consumers' contributions.