Decision Model and Notation for Describing Variability in Business
Process Product Lines
Andreas Speck, Aljoscha Jagenow and Melanie Windrich
Institute of Computer Science, Christian-Albrechts-University, Kiel, Germany
Keywords:
Decision Model and Notation, Modeling Variability, Software Product Lines for Business Process Models.
Abstract:
Product line techniques have proven being a successful concept of software reuse. Hence, product line tech-
niques have been applied for business process models. Although there are well established models for de-
scribing business processes such as BPMN or ARIS there is a lack of describing the variability in business
processes.
In the paper, we propose to apply Decision Model and Notation (DMN) for modeling process variability. DMN
has the advantage that it is a known concept and well understood. DMN supports the decision techniques by
defining the variability in business processes and with the rules defining the variability.
1 INTRODUCTION
Software product lines support the reuse of software
components on a large scale. Software systems are
considered as set of reusable assets with specific rela-
tions among them. The configuration of the assets is
variable. The variability of assets in software product
lines is an important issue in modeling. This requires
a comprehensive notation. (Svahnberg et al., 2001)
presents aspects of features and their modeling.
Based on the successful early approach, the com-
bination of software product lines and business pro-
cesses was discussed. Business processes are key el-
ements in defining and describing the sequences of
activities / workflows mainly in a commercial or ad-
ministration context. Such software systems like en-
terprise resource planning (ERP), e-commerce or ad-
ministrative systems are the organizational backbone
of most businesses or administrative units. Business
process models are used to define the behavior of such
systems (dos Santos Rocha et al., 2015). Due to the
high importance of business process model systems it
is very desirable to find techniques making such sys-
tems more reusable like the product line concepts.
A summary and survey about approaches combin-
ing business process models and product lines may
be found in (dos Santos Rocha and Fantinato, 2013).
A conclusion of this review is that the approaches of
applying software product line techniques to business
process models is very promising. However, there is
further need for improvement. A specific weak point
is the modeling of the variability in business process
models. This need of improved modeling of variabil-
ity is pointed out in more recent surveys like (La Rosa
et al., 2017) and (Pol’la et al., 2020). This demands
a modeling concept for variability in business pro-
cesses.
In the paper, we focus on modeling variability in
business process product lines. We propose to use al-
ready existing and established model types for vari-
ability modeling. An analogous approach proven al-
ready as successful in product line modeling in which
UML models are using UML stereotypes (La Rosa
et al., 2017). First approaches introducing stereotypes
are for instance (Clauß, 2001) or (Speck et al., 2002).
In this paper, we propose to use BPMN (Business Pro-
cess Model and Notation) as a basic notation for de-
scribing the processes which are the invariant parts of
the product lines. This is an attempt taken by a num-
ber of approaches applying product line techniques to
business processes (La Rosa et al., 2017; Pol’la et al.,
2020).
For describing variability in business process
models we propose to look on Decision Model and
Notation (DMN). As a tool, we use the CAMUNDA
modeler. However, our approach is independent from
a specific tool.
In this paper, we first introduce the background of
business process models, product line modeling and
DMN. Then we show how business process variabil-
ity may be modeled by DMN which is followed by an
example in the domain of e-commerce systems which
Speck, A., Jagenow, A. and Windrich, M.
Decision Model and Notation for Describing Variability in Business Process Product Lines.
DOI: 10.5220/0010497204450452
In Proceedings of the 16th International Conference on Evaluation of Novel Approaches to Software Engineering (ENASE 2021), pages 445-452
ISBN: 978-989-758-508-1
Copyright
c
2021 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
445
Figure 1: FODA Product Line Model for an Order Process.
demonstrates how to apply this approach.
2 BASIC CONCEPTS OF
PRODUCT LINES AND
BUSINESS PROCESS MODELS
At first, we have a look how variability is modeled
in product line systems. The we present the elements
of the Business Process and Model Notation (BPMN)
used in the paper and introduce then the Decision
Model and Notation (DMN).
2.1 Variability in Product Lines
The Feature-Oriented Domain Analysis (FODA)
(Kang et al., 1990) is a valid model for describing
dependencies between features and feature variabil-
ity. An example of a FODA model representing an
order process in an e-commerce system is depicted
in Figure 1. FODA models arrange the features in
tree structure (in our example Approval, Transaction
and Fulfillment are child features of the Order Pro-
cess). There are three different types of dependencies
between the parent feature and its child features (or
sub-features):
The mandatory dependency requires a child fea-
ture.
A child feature may be optional.
Alternative dependencies requires the selection of
one of the child features.
However, systems may rarely be modeled in a pure
tree structure. In most cases the systems are orga-
nized in a net structure. The cross tree constraints
(REQUIRES and MUTEX ) of FODA express rela-
tions between the features in different branches. Two
of these cross tree constrains in the example are:
shipping cost requires shipping and
pay on delivery and electronic delivery are mutu-
ally excluded.
2.2 Business Process and Model
Notation
The Business Process Model and Notation (BPMN)
supports the graphical modeling of business pro-
cesses. The main elements comprise tasks (repre-
senting the basic functionality of the process), work-
flow objects (flow objects such as different gateway
types or events or connection objects representing
the sequence flow). Pools and swim lanes orga-
nize and identify major participants. Further artifacts
are data objects or annotations. A pure example of
a BPMN diagram with some important BPMN ele-
ments is shown in Figure 2 (left side). The tool used
to model this diagram is Camunda. (BPMN, 2021a)
is a Camunda BPMN documentation.
The BPMN diagram (on the left side of Figure 2)
shows a business process model which may serve as
a base for product lines. The basic mechanisms to
realize variability in process models are the attach-
ment or detachment of sequences to the process mod-
els (Sinnhofer et al., 2015; La Rosa et al., 2017). This
means that the paths in the process model may be ex-
tended or shortened. BPMN supports the modeling
of variability. Branching operators of the gateways
such as parallel and exclusive may be sufficient for
describing options and alternatives in a process model
which are important for product lines concepts. How-
ever, purely using these operators will hamper the
ease of understanding business process product line
model and will lead to confusion. The question is now
how to express the variability beyond the pure usage
of gateways as branching operators. The description
of variability need to be separated from the process
ENASE 2021 - 16th International Conference on Evaluation of Novel Approaches to Software Engineering
446
Figure 2: Exemplary BPMN process model (left) and DMN decision requirement diagram on the right.
models.
2.3 Decision Model and Notation
Decision Model and Notation (DMN) (OMG, 2021;
BPMN, 2021b) supports modeling variabilities and
decisions to be taken. There are diagrams, tables and
and the expression language (Friendly Enough Ex-
pression Language, FEEL). The decision diagram or
to be more precise decision requirements diagrams is
at the right side of Figure 2:
Decisions: several inputs (like input data or the
knowledge) determine an output of the decision.
Tables may be used to describe the decision logic.
Input data: Information necessary determining a
decision.
Knowledge Model : decision logic which may be
reused. The decisions based on the same knowl-
edge model may lead to different results depend-
ing on inputs or sub-decisions.
Knowledge Source: Authorities, committees, reg-
ulations or policies which influence the knowl-
edge model.
The decisions may be hierarchical. The dependencies
are represented by the directed edges.
The decision tables describe the logic of the de-
cision. The rules are expressed by the rows. The el-
ements of the rules (input and output) are kept in the
columns. The input and output within a rule are con-
nected. The decision tables are attached to the deci-
sions in the DRDs (decision requirements diagram).
Figure 5 in the application example (Section 4) shows
a decision table. The decisions are attached to the
business process models and the decisions represent
the behavior at the gateways.
In modeling the knowledge of the decisions is sep-
arated in the DMN and not kept directly in the busi-
ness process models.
3 VARIABILITY DEFINITION BY
DECISION MANAGEMENT
NOTATION
There are three elements for describing the variabil-
ity in business process models by DMN: decision re-
quirement diagrams (DRD), decision tables and the
variability in business process models itself.
The decisions in the DRD (introduced in previous
Section 2.3) may describe the variability in the pro-
cess. Like the FODA models it is suitable for arrang-
ing the decision requirement diagrams in a tree struc-
ture. This supports a hierarchical structure of variants
(or decisions) in which decisions may be composed to
form underlaying sub-decisions. In contrast to FODA
the DRD do not support the differentiation of types of
dependencies between parent and child features (or
decision and sub-decision). This may be solved by
the decision table. The DRD gives hints about the
motivation for the decisions:
Knowledge models represent the decision logic.
Knowledge sources are origins of knowledge
models. These may be authorities or regulations.
Input data may directly drive the decisions.
The cross-tree dependencies are mutex and re-
quires. Requires requests a specific output. Mutex
may be expressed by a combination of input and falsi-
fication of the output value. Similarly alternative and
optional rules may be expressed by the combination
of rows in the tables.
The activation of the rules is determined by the
Hit Policy: The policy Unique (used in our example
in Figure 5) means that only one rule matches. The
policy first means that the first rule is applied. Exam-
ples for further rules are any, rule orders, output order
or different types of rule collection. Examples of the
usage of rules with decision tables may be found in
documentations like (BPMN, 2021b; OMG, 2021).
Decision Model and Notation for Describing Variability in Business Process Product Lines
447
Order Process
choose
Payment
Method
Service Delivery
chosen
Electronic
Delivery chosen
Shipping chosen
Order Process
Initiated
choose
Fulfillment
Method
Order Process
successful
Approval
Payment
Order Process
completed
Pay by Bank
Transfer chosen
Pay by Credit
chosen
Pay by Delivery
chosen
Pay by Bill chosen
Service Delivery
Conditions
Shipping Data
e.g. Shipping Cost
Calculate Price
Get Price from
Price List
Price List
Service Delivery
chosen
Electronic
Delivery chosen
Shipping
chosen
Taxation
Add Shipping
Cost
Payment Fees
Pay by Credit
chosen
Add Credit Card
Fees
Add Pay by
Delivery Fees
Pay by Delivery
chosen
Pay by Bank
Transfer chosen
Pay by Bill chosen
Discount
approved
not approved
Decission for
Electronic
Delivery
Payment Method
Pay by Bank
Transfer
Pay by Credit
Card
Pay by Bank
Transfer chosen
Payment
Pay by Delivery
chosen
Fulfillment
Payment on
Delivery
Payment by Bill
Pay by Bill failed
Order Process
successful
Fulfillment
successful
Fulfillment Method
Pay by Bill chosen
Payment
successful
Pay by Delivery
chosen
Fulfillment
successful
Fulfillment
Shipping
Electronic
Download
Service Delivery
Shipping Data
e.g. Shipping Cost
Service Delivery
Conditions
Credit Card
Payment failed
Pay by Delivery
failed
Pay By Bank
Transfer failed
Decission for
Shipping
Decission for
Service Delivery
Pay by Bill chosen
already
payed
Approval
"false"
Taxation
"true"
Taxation
"false"
Discount
"false"
Discount
"true"
Approval
"true"
Shipping
Cost
"true"
Pay by
Bank
Transfer
"true"
Pay by Credit
Card chosen
Pay by
Credit
Card
"true"
Pay by Delivery
chosen
Pay on
Delivery
"true"
Pay
by Bill
chosen
Pay by
Bill
"true"
Shipping
Service
Delivery
Shipping
"true"
Electronic
Delivery
Electronic
Download
"true"
Service
Delivery
"true"
Figure 3: Example: BPMN Order Process Model.
ENASE 2021 - 16th International Conference on Evaluation of Novel Approaches to Software Engineering
448
4 EXAMPLE FOR APPLYING
DMN
The order process in e-commerce systems serves as
example in the paper. We already introduced this pro-
cess in Section 2) as example for the FODA variabil-
ity model (c.f. Figure 1). We now focus on the dy-
namic behavior modeled as a process. For compari-
son we use the similar functionality as in the FODA
example.
The example consists of a decision requirements
diagram (DRD), an exemplary decision table that are
separated from the BPMN process model in which the
decisions define the variability.
4.1 BPMN of the Order Process
The decisions presented in the subsequent subsections
affect the BPMN process. In the process these deci-
sions are taken and evaluated according to the rules in
the decision tables.
The Order Process BPMN is shown in Figure 3.
All branches with positive decisions are marked by
‘‘true’’.
The decision about the fulfillment method and the
payment method are made at the beginning of the pro-
cess. Each of these decisions result in events thrown
representing the respective decisions.
The Calculate Price sub-process first looks up the
concrete price in the Price List. This task is the only
mandatory task in the entire sub-process. The tasks
Taxation and Discount are based on the sub-decisions
respectively and use the decision tables receiving in-
put form their knowledge models. The Payment Fees
sub-process and the potential adding of Shipping Cost
are based on the events thrown by the tasks in which
these methods are chosen.
All the previously mentioned tasks are part of the
Order Process pool. The participants of this pool
take care of the order process. However, there are
other participants caring about the payment and the
fulfillment. In e-commerce systems payment and ful-
fillment services are performed by specific systems
called when required. Hence, these functionalities
are modeled in separate pools. The communication
between the processes is realized by signal events in-
stead of message events for a better overview since the
explicit connecting edges are not required (BPMN,
2021a).
The Payment Method process is divided in two
parts in between which the Fulfillment Method is pro-
cessed, since there is the possibility to pay before or
after the fulfillment. All the tasks in the Payment
Method and Fulfillment Method pools are depending
on the events thrown at the beginning of the Order
Process. The payment attempts may fail which is
considered. However, the concrete effects of a failed
payment are not modeled in this example. The de-
cision table for the Payment Method is described in
Section 4.3
4.2 DRD Example
The Order Process Product Line DRD in Figure 4
1
is considered as a decision which consists of the sub-
decisions Aproval, Calculate Price, Payment Method
and Fulfillment Method.
The (sub-)decision Approval is driven by the Com-
pliance Rules for Approval as business knowledge
which is imposed by the knowledge source of the
Compliance Rules Team which may be the legal ad-
visors of the company.
The (sub-)decision Calculate Price depends on
the rules of the Price List knowledge model. The sub-
decisions are Taxation, Discount , Payment Fees and
Shipping Cost. Taxation is driven by the knowledge
model Taxation Rules which are based on the source
of Legal Taxation Regulations. An example of such
a rule may be that for purchases from certain other
countries no value-added tax has to be withheld. If a
Discount is realized depends on the rule in the Dis-
count Decision Table. Two different types of fees
Credit Card Fees and Pay on Delivery Fees deter-
mine the Payment Fees decision. The Credit Card
Fees and Pay on Delivery Fees also have an impact on
the sub-decisions Pay by Credit Card or Pay on De-
livery as part of the Payment Method (sub-)decision.
The sub-division of Shipping Cost requires Shipping.
The (sub-)decision Payment Method uses one con-
crete kind of payment. The alternatives are the sub-
decisions Pay by Bank Transfer, Pay by Bank Credit
Card, Pay on Delivery and Pay by Bill. The concrete
realization of one of these sub-decisions is driven by
the input data Choose Payment Method. Pay on De-
livery interacts with Electronic Delivery. This mutual
exclusion is expressed explicitly in the decision table.
The sub-decisions Service Delivery, Electronic
Delivery and Shipping form the Fulfillment Method
(sub-)decision.
4.3 Exemplary Decision Table
We take the Payment Method decision table as de-
picted in Figure 5. This decision table is directly affil-
1
The figure is not a screenshot of CAMUNDA but re-
drawn for better visibility. The DRD model of Camunda is
rather voluminous.
Decision Model and Notation for Describing Variability in Business Process Product Lines
449
Figure 4: Decision Requirements Diagram (DRD) for the Order Process Model.
iated to the Payment Method decision and defines the
decisions.
The table consists of four rules, each in a row. All
rules depend on the input of the Payment Method cho-
sen. These are the already presented Pay by Bank
Transfer, Pay by Credit Card, Pay on Delivery and
Pay by Bill. These will result in the corresponding
events with similar names, Payment Fees when there
are such fees and a determination of the Fulfillment
Method. In case of Pay on Delivery there is only Ship-
ping permitted.
Such decision tables may exist for the other
(sub-) decisions like Calculate Price and Fulfillment
Method. Since the rules in these tables will be in gen-
eral similar, we do not present them.
5 RELATED WORK
In general the concept of reuse is to some extend sup-
ported by business process models, however, reuse is
not the main focus. Reuse has been considered in
reusing certain functions (Becker et al., 2000). Com-
plex systems like product lines have not been consid-
ered.
With the orchestration of web services the reuse
on a large scale emerged. Not only single services are
to be reused, but also entire process sequences. An
example for typical first approaches to apply product
line concepts for business processes for web services
is (Fantinato et al., 2010). The Business Process Ex-
ecution Language (BPEL) is the modeling language
for orchestration web service systems. As micro ser-
vices may be considered as successors based on web
services micro services are also subject of reuse con-
cepts (Sinnhofer et al., 2020).
A concept of adapting business processes in
variant-rich workflow-based systems is using the con-
text in a certain state for determining which variant of
the process is to be realized. The variants need to have
a context-aware configuration. (Saidani et al., 2015)
presents a concept of context awareness for adapting
business process models. This context awareness is
achieved by a context meta-model supported by an
ontology. This adaptation of business process models
is some kind of variability. Business objects are con-
sidered as long-living reusable assets while the pro-
cesses are refined (Rulle and Siegeris, 2014).
A recent overview of the modifications in busi-
ness process models in order to achieve variability is
presented in (La Rosa et al., 2017). All approaches
share the same basic concept of extending or short-
ening process paths in the same way as presented in
this paper. Feature trees are also used by many ap-
proaches. However, these feature trees are used in the
traditional way like FODA (c.f. Figure 1). Although
BPMN as process model notation is used by many ap-
proaches, none of these previous approaches proposes
to use decision model and notation (DMN).
A remarkable idea is supporting the distributed
product line configuration by web-based systems.
An example is ProductlinRE, an online management
tool for requirements engineering of software prod-
uct lines (Ghofrani and Fehlhaber, 2018). The on-
ENASE 2021 - 16th International Conference on Evaluation of Novel Approaches to Software Engineering
450
Figure 5: Decision Table for Payment Method.
line system uses feature trees describing the variabil-
ity. From different locations these feature tree models
may be accessed and the variants may be selected in
teamwork. Such distributed working would also be
supported by using decision requirements diagrams
(DRD) which also bear the option to be used in a dis-
tributed environment.
Some of the features in a decision requirements di-
agram (DRD) may be cross-cutting in the meaning of
aspect-oriented programming (AOP) or separation of
concerns (SoC) concepts. Applying aspect-oriented
concepts allows to identify the cross-cutting features
with dependencies across the entire feature model
(Varela et al., 2011). These cross-cutting features may
be considered as cross tree constraints. The advantage
of clustering features as aspects and separating them
in concerns is that these are more easy to handle. And
entire groups of features may be activated or deacti-
vated in one step. Such an approach may be supported
decision requirements diagrams (DRD) and decision
tables. However, a specific notation for expressing the
aspects does not exist yet.
The basic understanding of the capabilities of De-
cision Management and Notation (DMN) may be
found in the standard documentation (OMG, 2021) or
basic tutorials like (BPMN, 2021b). (Taylor et al.,
2013) explains the background for applying DMN.
(Figl et al., 2018) presents further examples for ap-
plying DMN, e.g. hierarchies of decisions, the separa-
tion of decisions and process models or conventions.
These are base of our work.
6 CONCLUSION AND FUTURE
WORK
Software product lines are well established concepts
supporting software reuse on a large scale. Busi-
ness process models are a typical method for de-
scribing the behavior of commercial software systems
like ERP (enterprise resource planing) systems, e-
commerce or administrative systems. Due to the im-
portance of both product lines and business process
models it is an obvious choice to seek for a combi-
nation. As an extension of the well established busi-
ness process model and notation (BPMN) the concept
of decision model and notation (DMN) has emerged.
DMN may help to solve the lack of appropriate mod-
eling languages for variabilities (as desired by the
product line concept) in business process models and
to reduce the complexity of the model as proposed in
(Taylor et al., 2013).
We propose applying DMN models and decision
tables as means to define the variabilities in business
process product lines. The variability in product lines
are typically modeled in a tree structure. Decision
requirements diagram (DRD) provide a similar tree
structure in which the variability of the features which
in DMN are called decisions (resulting in functions
which realize the features) may be defined. In contrast
to traditional feature tree structures DRD also provide
the model elements representing the origin of the de-
cision rules. The rules themselfes are defined in the
decision tables. In rule tables both types of depen-
dencies (mutual exclusion and require) between deci-
sions may be expressed. The concrete location of the
variants in the business process model are defined in
the process model (in BPMN notation).
In our current approach, we do not consider yet
another approach to express variants besides DMN:
Case Management Model and Notation (CMMN).
CMMN provides modeling elements which reduce
the effort for describing the variants in a process
model. The models provide a more condensed view
of the variants. The open issue is if such simplified
modeling will be accepted be the systems developers
and if CMMN provides all the required techniques for
modeling the variants.
Further future challenges are techniques for auto-
Decision Model and Notation for Describing Variability in Business Process Product Lines
451
mated checking in the phase of the development of a
business process product line. This comprises the de-
velopment of the process model itself as well as the
design of the variants. The consistency of the variants
and correctness of order in the process models need
to be assured. An automated support of this quality
control would be desirable.
REFERENCES
Becker, J., Rosemann, M., and von Uthmann, C. (2000).
Guidelines of business process modeling. In Busi-
ness Process Management, Models, Techniques, and
Empirical Studies, volume 1806 of Lecture Notes in
Computer Science, pages 30–49. Springer.
BPMN (2021a). BPMN Camunda, Documentation,
https://camunda.com/bpmn/.
BPMN (2021b). Camunda, DMN Tutorial,
https://camunda.com/dmn/.
Clauß, M. (2001). A proposal for uniform abstract mod-
eling of feature interactions in UML. In Proceedings
of the ECOOP 2001 Workshop on Feature Interaction
in Composed Systems (FICS 2001), Budapest, Hun-
gary, June 18-22, 2001, volume 2001-14 of Technical
Report, pages 21–25. University of Karlsruhe, IPD.
dos Santos Rocha, R. and Fantinato, M. (2013). The use of
software product lines for business process manage-
ment: A systematic literature review. Information and
Software Technology, 55(8):1355–1373.
dos Santos Rocha, R., Fantinato, M., Thom, L. H., and Eler,
M. M. (2015). Dynamic product line for Business
Process Management. Business Process Management
Journal, 21(6):1224–1256.
Fantinato, M., de Souza Gimenes, I. M., and de Toledo,
M. B. F. (2010). Product Line in the Business Pro-
cess Management Domain, chapter 20, pages 497–
530. Taylor and Francis Group, LLC.
Figl, K., Mendling, J., Tokdemir, G., and Vanthienen, J.
(2018). What we know and what we do not know
about DMN. Enterprise Modeling and Information
Systems Architectures EMISA Forum, 13(2):24–26.
Ghofrani, J. and Fehlhaber, A. L. (2018). Productlinre:
online management tool for requirements engineer-
ing of software product lines. In Proceeedings of
the 22nd International Systems and Software Product
Line Conference - Volume 2, SPLC 2018, Gothenburg,
Sweden, September 10-14, 2018, pages 17–22. ACM.
Kang, K., Cohen, S., Hess, J., Novak, W., and Peterson, A.
(1990). Feature-Oriented Domain Analysis (FODA)
Feasibility Study. Technical Report CMU/SEI-90-TR-
2, Carnegie Mellon University, Software Engineering
Institute, Pittsburgh, PA.
La Rosa, M., van der Aalst, W. M., Dumas, M., and Milani,
F. P. (2017). Business Process Variability Modeling:
A Survey Article. ACM Computing Surveys, 50(2):1–
45.
OMG (2021). OMG, Decision Model and Notation (DMN),
https://www.omg.org/dmn/.
Pol’la, M., Buccella, A., and Cechich, A. (2020). A. Anal-
ysis of variability models: a systematic literature re-
view. Software and Systems Modeling, pages 1619–
1374.
Rulle, A. and Siegeris, J. (2014). From a family of state-
centric PAIS to a configurable and parameterized busi-
ness process architecture. In Business Process Man-
agement - 12th International Conference, BPM 2014,
Haifa, Israel, September 7-11, 2014. Proceedings,
volume 8659 of Lecture Notes in Computer Science,
pages 333–348. Springer.
Saidani, O., Rolland, C., and Nurcan, S. (2015). Towards
a generic context model for BPM. In 48th Hawaii
International Conference on System Sciences, HICSS
2015, Kauai, Hawaii, USA, January 5-8, 2015, pages
4120–4129. IEEE Computer Society.
Sinnhofer, A. D., Oberhauser, R., and Steger, C. (2020).
From adaptive business processes to orchestrated mi-
croflows. In Proceedings of Business Modeling
and Software Design. BMSD 2020, pages 152–168.
Springer, LNCS.
Sinnhofer, A. D., P
¨
uhringer, P., and Kreiner, C. (2015).
varbpm - a product line for creating business process
model variants. In Proceedings of the Fifth Interna-
tional Symposium on Business Modeling and Software
Design - Volume 1: BMSD, pages 184–191. INSTICC,
SciTePress.
Speck, A., Clauß, M., and Franczyk, B. (2002). Concerns
of variability in ”bottom-up” product-lines. In In:
Proceedings of Second Workshop on Aspect-Oriented
Software Development, pages 19–24. GI, Gesellschaft
f
¨
ur Informatik.
Svahnberg, M., van Gurp, J., and Jan, B. (2001). On the no-
tion of variability in software product lines. Technical
report, University of Karlskrona/Ronneby.
Taylor, J., Fish, A., Vanthienen, J., and Vincent, P. (2013).
Emerging Standards in Decision Modeling—an Intro-
duction to Decision Model & Notation, pages 133–
146. WfMC, Workflow Management Coalition.
Varela, P., Ara
´
ujo, J., Brito, I. S., and Moreira, A. (2011).
Aspect-oriented analysis for software product lines
requirements engineering. In Proceedings of the
2011 ACM Symposium on Applied Computing (SAC),
March 21 - 24, 2011, pages 667–674. ACM.
ENASE 2021 - 16th International Conference on Evaluation of Novel Approaches to Software Engineering
452