A Meta-Design Approach for Collabora-
tive Process Modeling
Selim Erol, Felix Mo¨dritscher, Gustaf Neumann
Institute for Information Systems and New Media, WU Vienna
serol@wu.ac.at, fmoedrit@wu.ac.at, neumann@wu.ac.at
Abstract. Modeling business processes has become a complex design task due to a in-
creasing rate of organizational change. Therefore business process modeling is in search
for ways to reach beyond the rigidity of traditional approaches. In order to overcome new
challenges, like the involvement of a diverse stakeholder community and variability of mod-
els, end-user development (EUD) and collaboration are considered promising instruments.
In this paper we propose to enhance collaborative process modeling by EUD concepts.
Precisely we apply the Meta-Design guidelines to the domain of business process mod-
eling and present a first prototypic environment for collaborative process modeling along
with a case-study. The paper concludes by highlighting benefits and unsolved issues of
our Meta-Design based process modeling approach and by indicating the next steps of our
research.
Introduction
The variability of business processes and associated process models is a frequently
stated challenge in designing flexible and adaptive business process management
(BPM) systems (van der Aalst et al., 2003; Kettinger and Guha, 1997; van der
Aalst and Jablonski, 2000; Schonenberg et al., 2008; Dadam and Reichert, 2009;
Rosemann et al., 2008). While the research focus has mainly been concentrated on
creating flexible process modeling techniques and workflow management systems
little attention has been dedicated to an in-depth investigation on the flexibility and
openness of modeling environments, especially to enable end-user participation in
process modeling (Dadam and Reichert, 2009).
46
Proceedings of the 2nd International Workshop on Open Design Spaces (ODS 2010)
International Reports on Socio-Informatics, vol. 7, no. 2 (2010), pp. 46-62

Traditionally the process of process modeling follows a linear model with sharply
separated phases reaching from analysis to design of a process model and it’s in-
stantiation (Weske, 2007). Usually the modeling process is limited to design-time
of a system mostly being part of an ex-ante requirements elicitation process. Recent
research streams challenge this approach in various ways. For example Dadam and
Reichert (2009) have shown approaches for a run-time extension or modification
of process model instances which can be fed back to the original process model.
Thus, a shift from linear models to spiral models or incremental approaches can be
observed. The process of modeling can be seen as a continuous process involving
multiple stakeholders and has to satisfy multiple perspectives which might as well
change over time. Although not yet sufficiently addressed in research the impor-
tance of stakeholder participation has been stated in several empirical studies like
Davies et al. (2006).
Organizational context however requires tools which support a flexible and easy
participation in the modeling process and which are themselves flexible enough to
be adapted to a changing environment. Hereby, the emerging paradigm of end-
user development (Lieberman et al., 2006) and research initiatives in the field of
collaborative design offer promising approaches for the engineering of open model-
ing environments to foster the integration of diverse stakeholders into the modeling
process.
In this paper we argue that through flexible and open modeling environments
the effective involvement of end-users into the modeling process can be supported.
As a starting point for our research we articulate challenges of collaborative pro-
cess modeling and use a Meta-Design approach suggested by Fischer and Giaccardi
(2006) to derive key features for a respective modeling environment. We present
a preliminary wiki-based prototype for user-driven collaborative process modeling
and a case-study. Finally, we conclude by discussing our experiences and open
issues of our research work.
Challenges of collaborative process modeling
As a starting point to depict the challenges experienced in collaborative process
modeling we will refer to an adapted life-cycle model of a process model frequently
proposed in literature (van der Aalst and van Hee, 2004; Weske, 2007). The life-
cycle model reveals typical stages of a process model reaching from an initial design
idea to a more or less formalized process model.
The life-cycle of a process model is determined by the purpose of modeling
(Becker et al., 2000). One predominant purpose is the documentation of “as-is”
processes to identify shortcomings and potential improvements. As-is models also
serve as a knowledge base for participants in a business process. In many cases it
is important to model a “to-be” state of a process. To-be models help businesses to
understand organizational and financial impact of a new process or process change
(Speck and Schnetgo¨ke, 2003). Another purpose is the creation of models that can
be used as a basis for process simulation or can be transformed into an executable
47

model description (mostly referred as workflow). Hereby, a major challenge deals
with dissolving the sharp distinction between “to-be” and “as-is” models thus lead-
ing to process models which are close-to-reality at all time of business operation.
Figure 1. The process model life-cycle showing the different phases and related model representa-
tions.
Concerning the artifact as the outcome of collaborative modeling such a process
model is characterized by complexity and volatility (Cardoso et al., 2006; Gruhn
and Laue, 2006). Complexity is caused by the number of variables that determine
a process. Process models are typically constituted by structural entities such as
activities, flow logic, events, conditions, decision nodes and attributes like costs
and resource allocation. Volatility arises as processes and therefore process models
have to be adapted to an increasing rate of changing conditions. A process model
artifact runs through a life-cycle starting with a tacit model and ending with a more
or less formalized and executable model description (see Figure 1). Respective
tools which claim to support the collaborative modeling process would have to en-
able the coexistence of multiple representations of a model ranging from informal
descriptions of a process (e.g. a narrative text), semi-structured descriptions (e.g.
a use-case template), graphical process models (e.g. a BPMN1 diagram) and ex-
1 Business Process Modeling Notation, for details see http://http://www.omg.org/bpmn/
48

ecutable descriptions (e.g. a BPEL2 description). A closely related challenge is
stated by Renger et al. (2008): namely the choice of a good starting point for a
collaborative modeling effort, meaning that usually an initial or preliminary model
provokes communication and problem structuring.
The process model life-cycle can also be analyzed with respect to the diversity
of stakeholders typically involved in the design of a process model. First we can
distinguish novices from experts regarding the knowledge about a specific business
domain. Second we distinguish along the hands-on skills regarding process model-
ing and associated tools. A third dimension arises when we consider the knowledge
about the design of software systems that rely on a more or less formalized busi-
ness process model. Fourth, stakeholders in a business process typically have a
specific organizational role which imposes a specific involvement or behavior in the
life-cycle of a business process. Organizational roles range from managerial roles
typically having a broad view on a process to employees focusing only on the pro-
cess fragments that relates to the tasks they are expected to accomplish. Modeling
environments should have not only to support but foster flexible and easy participa-
tion of diverse stakeholders along the process model life-cycle. A special case in the
spectrum of stakeholders is the role of a facilitator. The importance of a facilitator
is stated in several studies (e.g. den Hengst (2005); Vennix (1999)) mainly because
it is considered to be vital for reaching shared understanding among stakeholders
and transferring different views on a process into a valid model.
The promises of end-user development
End-user development (EUD), as depicted by Lieberman et al. (2006), aims at shift-
ing development tasks from programmers to end-users and thus evolving systems
from being “easy to use” to being “easy to develop”, primarily to enable users
to design software solutions which highly fit to their needs and perceptions. The
EUD paradigm has emerged in the field of software engineering and reaches be-
yond the claim of agile methods (Cockburn, 2007) and participatory design (Muller
and Kuhn, 1993) which involve users at design-time in the development process but
do not enable them to actively shape applications to their specific needs at run-time.
Although rooted in software engineering EUD shares many basic ideas with disci-
plines like CSCW (Lieberman et al., 2006) and recent phenomena like Web 2.0 or
Social Media (cf. the long tail of software by Kraus (2009)).
The basic assumption behind EUD shows commonalities with findings in Busi-
ness Process Management (BPM). It is the increasing frequency of change and di-
versity in life-cycles for individuals and for organizations as well as an environment
which need an increasing flexibility and adaptivity of technical systems claiming to
support the user in his activities. Fischer and Giaccardi (2006) state that “creating
the technical and social conditions for broad participation in design activities is as
important as creating the artifact itself”. The EUD paradigm tries to be applicable
2 Business Process Execution Language, for details see http://www.oasis-open.org
49

to any domain involving software support and does not anticipate much about the
outcome or object of a collaborative design process. In the case of business pro-
cess modeling the outcome of a socially enabled design process would be a process
model artifact that represents multiple views and does not claim to be in a final
state. Thus, process model evolution is considered a natural and ever emergent phe-
nomenon which has to be included by design in any processing modeling technique
or tool.
The application of the EUD paradigm to process modeling environments seems
– at least from our point of view – to have potential for future research. In this
paper we investigate the potentials of the EUD paradigm and in particular the Meta-
Design approach by (Fischer et al., 2004) in the context of business process design.
A Meta-Design approach towards process modeling
In order to examine collaborative process modeling under the perspective of EUD,
we investigate the Meta-Design guidelines by Fischer et al. (2009). Basically, Meta-
Design comprises a set of very generic guidelines which are valuable for developing
and providing EUD environments. Due to our focus on process modeling we high-
light the important aspects of the Meta-Design guidelines and derive key features
for a collaborative process modeling environment according to the specific problem
domain.
Addressing design and usage in the scope of EUD (Lieberman et al., 2006), it
has to be manifested that modeling itself is considered as the design phase while ap-
plying models in business operation or instantiation of workflows in software com-
prises the usage phase. However, the border between design and usage is very in-
distinct, as collaboration between different users fosters adaptation of a pre-defined
environment during usage.
Moreover and with respect to the “seeding, evolutionary growth, reseeding”
(SER) model by Fischer and Giaccardi (2006), the process models do not have to be
complete before they can be used (e.g. filed in an archive, inspected by other users,
verified by experts, shared within a company or community, etc). For a collabora-
tive approach towards process modeling we even consider that partially incomplete
process definitions can be executed in a workflow engine and be complemented by
end-users (with different roles) on execution time (Neumann and Erol, 2009).
Support human-problem interaction
The very first Meta-Design guideline deals with the issue that domain experts do
not want to be bothered with computers and software problems but with domain-
specific tasks and challenges. Due to the complexity of computer systems, process
modeling tools and process models, an environment for designing and managing
process models must hide non-modeling issues (like operating system facilities,
distracting features or disturbing information) from end-users.
50

Therefore we argue, an environment for modeling business processes collab-
oratively must provide (1) the facilities for process modeling, (2) support mecha-
nisms for single user interactions (i.e. recommendations of activities and activity
sequences for specific situations and problems), (3) visual feedback on the process
definition (e.g. highlighting control flow errors), (4) feedback on collaboration with
other users (e.g. conflicts through concurrent editing), as well as (5) a catalog of de-
sign solutions to typical problems (i.e. pre-defined templates for business processes
which could be shared within a company or community).
Collaborative modeling tools must somehow act as a facilitator to enable users
to focus on the problem instead of hassling with the tool and provide features to
mediate shared understanding.
Underdesign for emergent behavior
The second guideline of the Meta-Design framework addresses the support of a tool
throughout the modeling process and recommends to “underdesign” the modeling
artifacts to achieve emergence of behavior. It requires that a model artifact is not
delivered as a finished product, but allows users to solve parts of the overall problem
stepwise thus supporting the concepts of “hackability” and “remixability” of (user-
created) solutions.
In the scope of process modeling a tool is required that enables users to modify
parts of (probably large) process definitions, to verify and store these fragments of
a process separately, and also to share them with others. In sum, users should be
able to decompose the design problem (a process definition) into smaller design
elements and combine them with other elements which can be even given by other
users. Concerning reusing design elements, such a process modeling tool might
also consider data given from other users as an additional component for designing
processes.
Enable legitimate peripheral participation
As a consequence of user contributions to software (i.e. the overall environment and
process models), the third Meta-Design guideline refers to policies and procedures
for incorporating this user input for software tools and for making them aware of
their influence on the system.
For the context of collaborative process modeling, a tool requires facilities for
sharing outcomes of other users, like process models or parts of them. In combina-
tion with the last guideline, underdesign for emergent behavior, user participation
can also be fostered by allowing “incomplete” processes as well as process frag-
ments, so that users can slowly start to get into process modeling instead of being
confronted with large, complex process models and the creation of them. Finally
and regarding Fischer et al. (2009), process modeling should also support so-called
satellite communities, i.e. spaces for people in a certain domain or working on spe-
cific types of processes and process fragments which will be incorporated into the
overall environment when mature.
51

Share control
The forth guideline deals with user control within EUD environment with the partic-
ular goal to support different roles in the modeling process as well as in the process.
Therefore, the Meta-Design framework proposes to enable users to share control
within software systems, e.g. by granting access to artifacts they have created or by
authorizing others for certain actions.
Projecting this guideline onto collaborative process modeling, the environment
must involve the different stakeholders, as elaborated in the former section, and give
them adequate authority (control) according to their responsibility and role. Fol-
lowing the experiences from successful open source projects (Fischer et al., 2009),
granting authority attracts user who want to influence the EUD environment as well
as the outcomes (the process models).
Due to the collaborative character of our process modeling approach, we high-
light two important issues here. On the one hand and with respect to privacy issues
in online communities (Dwyer et al., 2007), users must have control over their data
and share this control on a fine-grained level. On the other hand, collaborative
authoring of documents (Borghoff and Schlichter, 2000) required facilities for con-
trolling edit operations in order to avoid conflicts between different users. Both
aspects are relevant for collaborative modeling of business processes.
Promote mutual learning and support
As users have different levels of skills and knowledge, this Meta-Design guideline
refers to knowledge sharing mechanisms that encourage users to learn from each
others. An EUD environment for process modeling could include tools like forums,
mailing lists, chat rooms, and other tools to exchange knowledge among peer users.
Current technologies like social networking platforms (Facebook, MySpace, etc.)
also apply recommendation strategies to support users in working with a system
and connecting them to peers.
All in all, an environment for process modeling requires strategies for support-
ing users in using the design facilities, finding relevant artifacts (e.g. templates
or fragments of process definition) and peers (e.g. expert in the same domain or
the owner of a relevant business process), and fostering practice sharing within a
community.
Reward and recognize contributions
Similarly to the last guideline, this one addresses the need to motivate users for
actively participating in the evaluation of the EUD environment, precisely by re-
warding and recognizing their contributions. As motivation of human beings can
depend on many (intrinsic and extrinsic) factors, Fischer et al. (2009) postulate that
many different strategies could be applied.
In the context of process modeling, contributions must be assigned to users.
Apart from simple benefits like optimization of time and effort for an individual
52

contributions must be recognizable. It is necessary that they can be found within
a software system, or that they are suggested actively by a platform (as observed
at many networking sites). Furthermore, a reward strategy such as explicit user
feedback (cf. Facebooks like button or the possibility to comment items) is highly
recommended. Finally also statistics about usage or linking of process models and
navigation facilities might increase the findability of user contributions.
Foster reflective communities
The last guideline given by Fischer et al. (2009) focus on the utilization of collec-
tive intelligence in order to solve complex design problems. Thus, EUD environ-
ments should include facilities for collaboration and communication, i.e. to create a
shared understanding among domain experts and to build and sustain a community
of end-user developers. This aspect is of particular relevance for process modeling,
as several stakeholders from different areas are involved into business processes.
In accordance with key features derived in former guidelines and functionality of
social networking platforms, we propose typical features like sharing, rating, tag-
ging, and commenting process models, enabling collaboration and communication
between users, recommending process templates and peer users, etc.
Derivation of key features
Summarizing this section, the left-hand side of Table I gives an overview of the
original Meta-Design guidelines which are kept very general and thus applicable
for nearly all kinds of EUD environments. Next to each guideline we highlight the
key features of a collaborative process modeling environment which relates to this
specific guideline.
In the next section we will present a first prototype of a process modeling envi-
ronment considering at least some of the key features derived from the Meta-Design
guidelines.
A Wiki-based modeling environment
In a recent research effort we incorporated EUD principles into a wiki framework to
achieve a prototypic open modeling environment. The basic wiki application was
built upon a well established and broadly used open-source community platform,
namely OpenACS (see Demetriou et al. (2006)). XoWiki – a wiki framework for
OpenACS based on the object-oriented, Tcl-based scripting language XoTcl – is
currently implemented in the context of numerous e-learning platforms (Neumann,
2007).
The wiki approach seems – at least from our point of view – to fulfill the guide-
lines proposed by the meta-design framework in various ways (see references to
Table I in brackets). First, the wiki system is by design an open environment that
53

Meta-Design guideline & key concepts by Fischer et
al. (2009)
Derivation of key features of a collaborative process mod-
eling environment
1. Support Human-Problem Interaction: avoid computer
and handling problems, focus on domains and tasks of
end-users
(a) consideration of process modelers and non-modelers, (b)
recommendations of possible process tasks, (c) visual feed-
back of process definition, (d) feedback on collaboration, (e)
provision of design solutions (e.g. templates)
2. Underdesign for Emergent Behavior: use simple
modeling components to be reused by users
(a) possibility to partition large process models, (b) separate
verification and storage of process fragments, (c) sharing of
fragments
3. Enable Legitimate Peripheral Participation: provide
policies and procedures for user participation, create
awareness of influence
(a) involvement of other users (e.g. assigning modeling task),
(b) allowance of incomplete processes to be completed by
others, (c) possibility of merging process fragments of others
into a process, (d) support of satellite communities (see 1.)
4. Share Control: support different roles in the modeling
and the process, grant access to artifacts
(a) access permissions for processes and fragments, (b)
preservation of privacy in the community, (c) concurrency
awareness and control to avoid editing conflicts
5. Promote Mutual Learning and Support: foster knowl-
edge sharing amongst users
(a) practice and knowledge sharing through recommending
process tasks, fragments, and peer actors, (b) typical social
networking functionality
6. Reward and Recognize Contributions: motivate end-
users to participate actively (multiple strategies)
(a) possibility to assign modeling tasks and visual feedback
on overall outcome (see 3. and 1.) and user contributions, (b)
(implicit or explicit) user feedback on process models (com-
ments, I like button of Facebook, etc.)
7. Foster Reflective Communities: use collective intelli-
gence to solve complex design problems
(a) facilities for collaboration and communication amongst
end-users, (b) typical functionality of social networking plat-
forms (sharing, rating, tagging, recommending, etc.; cf. 5.)
Table I. Specialization of the (generic) Meta-Design guidelines for collaborative process modeling.
facilitates broad participation in the creation and modification of model artifacts re-
gardless the characteristics of a specific stakeholder. Second, typical wiki features
like descriptive page identifiers, simple linking by page titles, on-the-fly creation of
pages that do not already exist and flat structure of content are a good starting point
for an open and end-user oriented collaborative environment.
XoWiki supports multiple representations of process models ranging from sim-
ple textual descriptions to use-case templates and graphical representations (see Fig-
ure 2). The ability to configure the user-interface and the diversity of representa-
tions of a model offers each stakeholder an appropriate access point to articulate
design problems and find solution suggestions. XoWiki can be configured to al-
low only strict formal process descriptions through forms and associated validation
mechanisms but also may be configured to allow simple (underdesigned) textual
descriptions and graphical diagrams (→3 1a, 3a). Process models and fragments
can be enriched with various media formats like documents, videos, images, audio
files, etc. Such rich process descriptions can be collaboratively developed, linked,
bookmarked, searched etc. in a single environment (→ 5b, 7b).
3 We use the right arrow symbol to symbolize references to the guidelines in Table I
54

Figure 2. Different user-interfaces representing a textual description of a simple sales process and
an associated graphical representation.
Process descriptions of any kind can be divided and linked into arbitrary frag-
ments (→ 2a, 2b, 3b, 3c). Vice-versa process model fragments can be connected to
form higher-level models. Apart from vertical fragmentation, horizontal fragmenta-
tion is possible by extending and linking fragments to each other. The combination
of bottom-up and top-down modeling is considered to support the creation of seeds
and emergent design behaviors. In addition to the mechanism for fragmentation
of process models legitimate peripheral participation is enabled via typical com-
mentary and tagging functions. Such process model descriptions can be exposed
to discussions, reflections and contributions from the community without neces-
sarily intervening the design process (→ 2c, 6b, 7a). The coexistence of different
descriptions each reflecting a specific view and state of a process model fosters
knowledge sharing and mutual learning among multiple stakeholders. The explo-
ration of comments and revision histories additionally enables stakeholders to learn
from the evolution of a process model description.
In scenarios (e.g. in organizations) where a certain degree of control of “who-
may-edit-what” is desired XoWiki offers a fine-grained user and role concept to de-
fine content policies from complete openness to multi-level access policies (→ 4a).
The content flow component of XoWiki (Neumann, 2008) provides explicit state
modeling for all wiki objects. By using this component, one can define workflows
for the processing of wiki pages. One can for example define release processes,
55

complex page flows or other user interaction sequences. The workflow is based on
the state design pattern (Gamma et al., 1995) and can therefore provide a different
behavior and/or different presentation depending on the state and the defined user
roles (→ 4a). This mechanism can also be used to define the development process
of a business process models. One can offer different user-interaction controls at
design time of the business process models, or define release steps, or simply make
the state of a modeling fragment explicit.
Through the concept of activity graphs (Neumann, 2007) the system provides
an awareness mechanism which reveals individuals contribution to a process model
(→ 1d, 5a, 6a).
The process modeling wiki described is fully implemented as a prototype that
can be used for a web-based shared development of business process models. So
far, we have not evaluated the system in the larger scale but we have conducted first
usability studies and a case-study which will be described in the next section.
Case-study: a book order process at New Media Lab
The Case
At the Institute of Information Systems and New Media at the University of Busi-
ness and Economics in Vienna we have recently decided to support our book or-
dering process by means of software. The basic idea was to enhance the existing
process which was characterized by multiple entry of the same data and mostly
outdated information on the book’s actual location and availability. This resulted
from the use of multiple applications (e-mail, local database application, library
catalog) to order a book and store data on a book. The administration of the book
records by office staff lead to an out-of-date information on the books’ actual loca-
tion and availability. This was mainly due to the fact that staff members exchanged
books without informing the office management. The meta-data (author, title, year
of publication, ..) of a book proved to be incorrect or incomplete in many cases.
The process knowledge which existed tacitly scattered among the different parties
involved was not externalized in any way.
In order to capture the different views and knowledge of the “as-is” process
we used the wiki-environment proposed above. Therefore we created some intro-
ductory pages and also linked to external resources, especially for tutorials on the
modeling technique to be used. The introductory pages can be classified in infor-
mation about the general use of the wiki, information about the purpose and goal of
the case-study and information on how to contribute to a process model. We also
included a sample process model and a preliminary and yet very high-level model
of the book order process.
Along with the invitation of the stakeholders we conducted a short individual
training to explain the background of the task and give some motivational support.
The stakeholders finally involved comprised five members of the scientific staff both
at senior and junior level, the head of the institute and our office management staff.
56

While scientific staff partly had expert modeling skills the employee responsible for
book administration did not have any knowledge about process modeling.
Lessons learned
The case-study presented reflects a typical situation in process modeling. The cap-
turing of the “as-is” process is the basis for a continuously evolving “to-be” process
model. In the following we will outline selected experiences we gained during this
study.
During our introductory face-to-face sessions at the beginning of the modeling
we frequently experienced the objection that this process is that simple that it can
be described in one sentence. We traced that back to the fact that each stakeholder
only had a limited view on the process and therefore initially underestimated the
complexity of the process. With the course of the study and the increasing number
of contributions these objections became obviously invalid.
Although the office management staff member had no modeling skills she could
quite accurately describe the sequence of activities, conditions and information ob-
jects involved in the process. To externalize her knowledge she used a simple wiki
page with a built-in rich-text editor. The representational style she has chosen was
narrative and semi-structured with bullet points. Her contribution proved to be a
valuable input for modeling experts.
During the starting phase reflections on the model to be created were mostly
dominated by concerns about how to use the tool. This valuable feedback for further
development of the tool somehow contradicts the first guideline of the meta-design
approach which demands that a respective tool should not hinder human-problem
interaction. However, the reflections on the tool abated with the further use of the
tool. Another issue that led to discussions during the design process was the usage
of symbols for modeling links to other process fragments or sub-process pages.
Some of the participants commented on the usefulness and usability of the tuto-
rials without recognizing the possibility to contribute or change them the way they
want them to be. We have to say that we did not stress this possibility explicitly
in the introductory session. We continuously modified and extended the learning
materials during the study according to user comments.
Although initially intended to have a high level process model developing into
detailed sub-process models we experienced a fragmentation taking place from the
very beginning. This was due to the fact that some of the modelers feared the inter-
ference with other modelers working on the same page. A first analysis of revision
histories of the model fragments revealed that some participants contributed many
revisions over several weeks. Others participated once contributing only one revi-
sion.
The role of a facilitator proved to be important in two ways: First, participants
needed a introduction to the task as a starting point for their contributions. Second, a
kind of linking and alignment had to be performed to integrate individual modeling
efforts.
57

The preliminary results of the case-study will enable us to reflect on the validity
of the guidelines proposed for collaborative modeling. For further evaluation we
will have to involve a broader audience into the modeling process (e.g. the company
responsible for shipping the product, library service) and observe the evolution of a
process model over a longer time-period.
Related work
Renger et al. (2008) summarized the most prevalent challenges of collaborative
modeling found in literature in the fields of problem structuring methods, group
model-building and enterprise analysis. One of the key issues is the importance of a
facilitator in the modeling process. Other key issues are the choice of a starting point
and modeling in parallel. The literature review is comprehensive but is focused on
providing a general (high-level) analysis of challenges in collaborative modeling.
Nonetheless, this served us as a basis for more concrete challenges especially when
it comes to requirements for tool support.
Recent empirical work was also conducted by Rittgen (2009, 2010). In several
experiments, case-studies and interviews he investigated business needs and success
factors for collaborative business process modeling. He studied practical business
experiences and needs in collaborative modeling. Although Rittgen’s studies pro-
vide a valuable basis for our approach it is actually limited to small and expert
oriented groups and does not provide insights when it comes to open modeling en-
vironments where the community of participants is not predetermined.
Decker et al. (2007) have investigated the use of wikis in requirements elici-
tation. They provide an outline of advantages in using wikis compared to other
tools. Many of the propositions are equally valid for collaborative process model-
ing. In (Decker et al., 2004) a methodological approach and platform is presented
for participative process modeling and learning. Results of three case-studies show
that higher user acceptance and perceived model quality can be reached through
user participation. A process modeling method especially designed for user par-
ticipation is suggested by Becker et al. (2007). The approach consists of a set of
language constructs and a procedure model which “allow for an easy and straight-
forward modeling of a public administrations process landscape”. The researchers
proved the applicability of their approach in several projects.
Concerning tool support for collaborative modeling we want to outline two dis-
tinct developments: one approach is presented by Hasso Plattner Institute and has
integrated a web-based modeling editor into Google Wave (Hasso Plattner Institute,
2010b; Dreiling, 2010). Thus, allowing users collaboratively annotate and com-
municate on process models while actively modeling. The same research group
has launched a community portal for sharing and collaboratively editing models
(Hasso Plattner Institute, 2010a). Another development is an open platform for
sharing scientific workflows (Roure et al., 2007).
58

Conclusions and outlook
In this paper we highlighted challenges and problematic aspects of collaborative
process modeling and proposed to apply concepts of end-user development to en-
hance process modeling. We also tried to come up with a EUD based approach
towards collaborative process modeling which is based on the Meta-Design guide-
lines. As a first proof-of-concept, we presented a prototypic environment which is
designed according to the Meta-Design guidelines for process modeling.
Overall we think that the complexity and volatility of designing business pro-
cesses requires such a step towards an open and collaborative design process. At this
early stage of our research it is already apparent that specific aspects – such as the
emergent behavior through under-designed, the active involvement of stakeholders,
scaffolding and mutual learning support, rewarding and reorganization mechanisms,
as well as reflection and community aspects – are of particular relevance for an open
process modeling environment. To underpin this assumption we have derived key
features for such modeling environments from the Meta-Design framework and in-
corporated them partly in a wiki-based modeling environment.
Although we have not realized all key features, we have developed a prototype
modeling environment that is intended to be an open design space where stake-
holders are encouraged to contribute and share knowledge regarding processes. For
instance, stakeholders are able to design process fragments within their scope of
responsibility and according to their domain knowledge. These specialized parts of
process models can be shared, reused and improved. Stakeholders are not limited to
one representational style of a process model (e.g. a process BPMN graph) but are
able to contribute in various ways (e.g. simple textual descriptions or comments).
The wiki-based modeling environment is not limited to support the continuous and
user-driven evolution of process models but allows also to collaboratively develop
associated artifacts like tutorials and other learning resources. Finally, the modeling
environment has also to be considered as subject of continuous evolution. Thus,
feedback mechanisms should not only be targeted at the design object – in this case
a process model – but also at the tool claiming to support the design process.
A case-study was conducted to validate the prototype modeling environment
along with the theoretical assumptions. The preliminary results of the case-study
show us that the users involved reflected both on the modeling object and the model-
ing environment. We experienced remarkable differences with regard to the extent
of contributions from individuals and the type of contribution (e.g. informal vs.
formal). During the case-study we also observed a strong need for instruction and
facilitation during the modeling process.
In sum, this leaves us two important issues to continue research work. On the
one hand, it is necessary to develop an evaluation framework for process modeling
environments which claim to support organizations in collaborative model build-
ing. Similar prototypical attempts from other research groups (see e.g. Hasso Plat-
tner Institute (2010b) and Roure et al. (2007)) encourage us to establish a framework
that provides basic theoretical concepts to collaborative modeling and furthermore
59

forms a basis for evaluation. On the other hand, we will go on with development
work on our wiki-based modeling environment and also evaluate it in the form of
user studies and ex-post data-analysis. We will have to rigorously study the impact
of different collaborative features on modeling outcome.
Acknowledgments
The research leading to these results has partly received funding from the Euro-
pean Community’s Seventh Framework Programme (FP7/2007-2013) under grant
agreement no 231396 (ROLE project).
References
Becker, J., L. Algermissen, D. Pfeiffer, and M. Ra¨ckers (2007): ‘Local, Participative Process
Modelling-The PICTURE-Approach’. In: Proceedings of the 1st Int. Workshop on Manage-
ment of Business Processes in Government at the 5th International Conference on BPM. pp.
33–48.
Becker, J., M. Rosemann, and C. Von Uthmann (2000): ‘Guidelines of Business Process Modeling’.
In: W. van der Aalst, J. Desel, and A. Oberweis (eds.): Business Process Management: Models,
Techniques and Empirical Studies., Vol. 1806 of LNCS. pp. 30–49, Springer Verlag.
Borghoff, U. and J. Schlichter (2000): Computer-Supported Cooperative Work: Introduction to
Distributed Applications. Springer Verlag.
Cardoso, J., J. Mendling, G. Neumann, and H. Reijers (2006): ‘A Discourse on Complexity of
Process Models’. In: Business Process Management Workshops 2006 LNCS 4103, Vol. 4103.
pp. 117–128, Springer Verlag.
Cockburn, A. (2007): Agile software development: the cooperative game. Addison-Wesley.
Dadam, P. and M. Reichert (2009): ‘The ADEPT Project: A Decade of Research and Development
for Robust and Flexible Process Support – Challenges and Achievements’. Computer Science -
Research and Development, vol. 23, no. 2, pp. 81–97.
Davies, I., P. Green, and M. Rosemann (2006): ‘How do practitioners use conceptual modeling in
practice?’. Journal of Data Knowledge Engineering, no. 58, pp. 358–380.
Decker, B., E. Ras, J. Rech, P. Jaubert, and M. Rieth (2007): ‘Wiki-based stakeholder participation
in requirements engineering’. IEEE SOFTWARE, pp. 28–35.
Decker, B., J. Rech, K. Althoff, A. Klotz, E. Leopold, and A. Voss (2004): ‘Participative process in-
troduction: Three case studies from the indiGo project’. Journal of Universal Computer Science,
vol. 10, no. 3, pp. 186–204.
Demetriou, N., S. Koch, and G. Neumann (2006): ‘The Development of the OpenACS Community’.
Open Source for Knowledge and Learning Management: Strategies Beyond Tools. Idea Group
Publishing.
den Hengst, M. (2005): ‘Collaborative Modeling of Processes: What Facilitation Support Does a
Group Need?’. In: AMCIS 2005 Proceedings. AIS.
60

Dreiling, A. (2010): ‘Gravity – Collaborative Business Process Mod-
elling within Google Wave, retrieved on July, 23rd 2010, from
http://www.sdn.sap.com/irj/scn/weblogs?blog=/pub/wlg/15618’.
Dwyer, C., S. Hiltz, and K. Passerini (2007): ‘Trust and privacy concern within social networking
sites: A comparison of Facebook and MySpace’. In: Proceedings of AMCIS.
Fischer, G. and E. Giaccardi (2006): ‘Meta-design: A framework for the Future of End-User Devel-
opment’. In: H. Lieberman, F. Paterno, and V. Wulf (eds.): End User Development. Springer
Verlag, Chapt. 19, pp. 427–457.
Fischer, G., E. Giaccardi, Y. Ye, A. Sutcliffe, and N. Mehandjiev (2004): ‘Meta-design: A Manifesto
for End-User Development’. Communications of the ACM, vol. 47, no. 9, pp. 33–37.
Fischer, G., K. Nakakoji, and Y. Ye (2009): ‘Metadesign: Guidelines for Supporting Domain Experts
in Software Development’. IEEE Software, vol. 26, no. 5, pp. 37–44.
Gamma, E., R. Helm, R. Johnson, and J. Vlissides (1995): Design Patterns: Elements of Reusable
Object-Oriented Software. Addison-Wesley Reading, MA.
Gruhn, V. and R. Laue (2006): ‘Complexity Metrics for Business Process Models’. In: 9th interna-
tional conference on business information systems (BIS 2006), Vol. 85. pp. 1–12.
Hasso Plattner Institute (2010a): ‘BPMN Community, retrieved on July, 23rd 2010, from
http://en.bpmn-community.org’.
Hasso Plattner Institute (2010b): ‘processWave, retrieved on July, 23rd 2010, from
http://www.processwave.org’.
Kettinger, W. and S. Guha (1997): ‘Business process change: a study of methodologies, techniques,
and tools’. MIS Quarterly, vol. 21, no. 1, pp. 55–80.
Kraus, J. (2009): ‘The long tail of software’. http://bnoopy.typepad.com/ (2009-04-02).
Lieberman, H., F. Paterno, M. Klann, and V. Wulf (2006): ‘End-User Development: An Emerging
Paradigm’. In: H. Lieberman, F. Paterno, and V. Wulf (eds.): End User Development. Springer
Verlag, Chapt. 1, pp. 1–8.
Muller, M. and S. Kuhn (1993): ‘Participatory design’. Communications of the ACM, vol. 36, no. 6,
pp. 24–28.
Neumann, G. (2007): ‘XoWiki – Towards a Generic Tool for Web 2.0 Applications and Social
Software’. In: OpenACS and .LRN Spring Conference, International Conference and Workshops
on Community Based Environments. Vienna, Austria.
Neumann, G. (2008): ‘XoWiki Content Flow – From a Wiki to a Simple Workflow System’. In:
Proceedings of 7th OpenACS / DotLRN Conference. Valencia, Spain.
Neumann, G. and S. Erol (2009): ‘From a Social Wiki to a Social Workflow System’. In: D.
Ardegna, M. Mecella, and J. Yang (eds.): Business Process Management Workshops 2008,
Vol. 17 of LNBIP. pp. 698–708, Springer Verlag.
Renger, M., G. Kolfschoten, and G. Vreede (2008): Challenges in Collaborative Modeling: A Liter-
ature Review, Vol. 10 of LNBIP, pp. 61–77. Springer.
Rittgen, P. (2009): ‘Collaborative Modeling – A Design Science Approach’. In: Proceedings of the
42nd Hawaii International Conference on System Sciences.
61

Rittgen, P. (2010): ‘Success Factors of e-Collaboration in Business Process Modeling’. In: Advanced
Information Systems Engineering. pp. 24–37, Springer.
Rosemann, M., J. Recker, and C. Flender (2008): ‘Contextualisation of business processes’. Inter-
national Journal of Business Process Integration and Management, vol. 3, no. 1, pp. 47–60.
Roure, D. D., C. Goble, and R. Stevens (2007): ‘Designing the myexperiment virtual research en-
vironment for the social sharing of workflows’. Future Generation Computer Systems, pp. 603–
610.
Schonenberg, H., R. Mans, N. Russell, N. Mulyar, and W. van der Aalst (2008): ‘Process Flexibility:
a Survey of Contemporary Approaches’. In: Advances in Enterprise Engineering I: 1st Inter-
national Workshop Ciao! and 4th International Workshop Eomas at Caise 2008, Montpellier,
France, June 16-17, 2008, Proceedings. p. 16.
Speck, M. and N. Schnetgo¨ke (2003): ‘To-be Modelling and Process Optimization’. In: J. Becker,
M. Kugeler, and M. Rosemann (eds.): Process Management – A Guide for the Design of Business
Processes. Springer Verlag, pp. 135–164.
van der Aalst, W., A. Hofstede, and M. Weske (2003): ‘Business Process Management: A Survey’.
In: LNCS – Business Process Management, Vol. 2678. p. 1019, Springer Berlin / Heidelberg.
van der Aalst, W. and S. Jablonski (2000): ‘Dealing with Workflow Change: Identification of Issues
and Solutions’. Computer Systems Science and Engineering, vol. 15, no. 5, pp. 267–276.
van der Aalst, W. and K. van Hee (2004): Workflow Management: Models, Methods, and Systems
(Cooperative InformationSystems). The MIT Press.
Vennix, J. (1999): ‘Group model-building: tackling messy problems’. System Dynamics Review,
vol. 15, no. 4, pp. 379–401.
Weske, M. (2007): Business Process Management – Concepts, Languages, Architectures. Springer
Verlag.
62