Information Technology Reference
In-Depth Information
Validation Exemplified. We demonstrate the validation of one map-
ping for an easy example from the case study excerpt of Fig. 1. We as-
sume a mapping from the subprocess W ishList to the mandatory feature
BusinessM anagement . The mapping is represented in Σ by an axiom
Map W ishList
feature.BusinessM anagement . We expect no con-
straint violations since the feature BusinessM anagement is mandatory.
Concerning the validation, f Φ
is build using the parent definition, i.e.,
f Φ ≡∃
hasF eature.StoreF ront
hasF eature.BusinessM anagement .
The expression f Ω
is build using the class Map W ishList
from the mapping
model ( f Ω ≡∃
feature.BusinessM anagement ). For the subsumption checking
of f Φ by f Ω , we can replace the property feature by hasF eature .Itiseasy
to see that due to the negation of f Φ , f Ψ
is equivalent to the top class
:
f Ψ ≡∀
hasF eature.
¬
StoreFront
hasF eature.
¬
BusinessM anagement
hasF eature.BusinessM anagement . In case a particular mapping causes a
violation, the user finds the corresponding constraint expression f Ψ
classified as
not equal to the top class (
).
Lessons Learned. In Sect. 3.3, we already distinguished the focus of our work
on well-formedness constraints to the work on behavioral profiles of [12]. After a
deeper comparison of both formalisms, we find two interesting aspects that di-
rectly impose further research challenges. Firstly, behavioral profiles are ecient
to compare process behavior and behavior consistency. It might be a promis-
ing step to extend our well-formedness constraint validation towards a behavior
constraint validation, while we still offer the same feature-oriented configuration
view in combination with the mappings. Secondly, in this context, we see poten-
tial on using DL for the validation. The main challenge from the DL modeling
perspective is to handle the possibility of concurrent executions of activities.
This problem seems to be in line with the descriptive modeling style of DL to
capture this kind of execution potentiality.
7 Related Work
Due to the increasing need of business processes customization, several ap-
proaches for the development of families of business processes have been in-
troduced like Schnieders et al. [13], Boffoli et al. [14], La Rosa et al. [15] and
van der Aalst et al. [16]. Schnieders et al. [13] model families of business process
models as a variant-rich business process model. A configuration of such a fam-
ily is performed by directly selecting business process elements of variant-rich
processes. In order to support such an approach, Schnieders et al. extend BPMN
with concepts for modeling variation. However, in order to perform it, such an
approach requires from a customer knowledge of business process modeling.
Boffoli et al. [14] and La Rosa et al. [15] also distinguish between business
process models and problem space models. Boffoli et al. model problem space as
variability table, while La Rosa et al. provide variability by questionnaires. They
provide guidance to derive valid configurations, while our aim is to guarantee
 
Search WWH ::




Custom Search