ARCHITECTURE FOR END USER-DRIVEN COMPOSITION OF
UNDERSPECIFIED, HUMAN-CENTRIC BUSINESS PROCESSES
Todor Stoitsev, Stefan Scheidl
SAP Research, SAP AG, Bleichstr. 8, Darmstadt, Germany
Felix Flentge, Max Mühlhäuser
Telecooperation Group, Darmstadt University of Technology, Darmstadt, Germany
Keywords: Software architectures, distributed applications, end user development, computer supported cooperative
work, ad-hoc workflow.
Abstract: Enterprises are constantly struggling to optimize their business processes in order to gain competitive
advantage and to survive in the fast evolving global market. Often, the only ones to understand the matter
and complexity of these processes are the people, who actually execute them. This raises the need for novel
business process management approaches, which can enable business users to proactively express process
knowledge and to participate in the business process management and design according to their actual
expertise and problem solving strategies. The presented paper describes an architecture, which supports a
framework for end user-driven composition and management of underspecified, human-centric business
processes. The solution builds up on email-integrated task management and enables dynamic generation of
decentralized-emerging process structures through web service-based activity tracking. The captured
process execution examples are shared in central enterprise repositories for further adaptation and reuse.
This enables “seeding, evolutionary growth, and reseeding” of user-defined, weakly-structured process
models towards global best-practice definitions.
1 INTRODUCTION
A significant shift in the view on enterprise
processes has been recognized in the last years.
While conventional workflow solutions are well
suited for static, predefined processes, the
importance of unstructured, knowledge-intensive
work in enterprises raises new flexibility
expectations (Aalst et al., 1999; Schwarz et al.,
2001). This is accompanied with the increasing
demand to enable business users to develop and
exchange substantial business process knowledge,
which could increase the overall enterprise
efficiency (Wiig, 2004). Recent analyst reports
clearly revealed that the traditional enterprise
process modeling perspective is being replaced by
tailoring of business processes according to the
individual point of view and connecting them
towards the achievement of common enterprise
goals (Gartner, 2006). This novel view raises new
challenges for the next generation Business Process
Management (BPM) by stating the fundamental
need to leverage individual expertise of business
users towards the definition and management of
agile business processes.
In the literature End User Development (EUD) is
defined as “a set of methods, techniques, and tools
that allow users of software systems, who are acting
as non-professional software developers, at some
point to create, modify, or extend a software
artifact.” (Lieberman et al., 2006). Within the
presented paper a process model is considered as a
software artifact, which can be enacted and reused to
support underspecified, human-centric processes.
The presented paper is motivated through the
possibility to “render” appropriation of process
models to end users and to “exploit the potential of
opportunity-based and emergent changes” from the
introduction of groupware in enterprises (Wulf &
Jarke, 2004). The paper describes a generic
architecture for enabling EUD of decentralized-
emerging, weakly-structured process models. The
165
Stoitsev T., Scheidl S., Flentge F. and Mühlhäuser M. (2008).
ARCHITECTURE FOR END USER-DRIVEN COMPOSITION OF UNDERSPECIFIED, HUMAN-CENTRIC BUSINESS PROCESSES.
In Proceedings of the Tenth International Conference on Enterprise Information Systems - DISI, pages 165-172
DOI: 10.5220/0001689501650172
Copyright
c
SciTePress
architecture is implemented in the Collaborative
Task Manager (CTM) prototype. The latter is not
explicitly discussed in the paper whereas references
to certain functionalities are mentioned as
clarifications for the presented concepts.
The paper is organized as follows. Section 2
presents related work in the area of agile process
support. Section 3 provides an overview of a
framework for light-weight composition and
management of ad-hoc business processes,
supported through the presented architecture. The
architecture is described in section 4. Conclusions
and future research directions are given in section 5.
2 RELATED WORK
Riss et al. (2005) discuss the challenges for the next
generation BPM by proposing the reuse of emerging
“task patterns” and “process patterns” as alternative
to static workflows. A further task-centric approach
which enables Proactive Information Delivery (PID)
on tasks and instance-based task reuse is presented
by Holz et al. (2005). Task centric approaches
bridging routine and ad-hoc work are also known
(Bernstein, 2000; Jorgensen, 2004). The above
studies present comprehensive strategies for
supporting knowledge work. However, they do not
consider methods or architectures for embedding ad-
hoc process support in the existing working
environment of end users. We suggest that this issue
is important as similarly to tailoring of software
systems, tailoring of ad-hoc processes through the
definition, adaptation and reuse of user-defined task
hierarchies should be ensured through “gentle slope
of complexity” (MacLean et al., 1990), where users
with different business and Information Technology
(IT) background can efficiently shape and exchange
reusable task definitions.
Evidences from related literature show that user
strategies for organizing daily activities are far from
any process or case-definition context and mostly
rely on common office tools such as email (Bellotti
et al., 2005) or personal to-do lists (Bellotti et al.
2004). Agostini et al. (1997) cross the boundaries of
the personal workspace and integrate to-do lists and
email within email-based workflows. However, the
authors do not discuss mechanisms for decoupling
email-based workflows from the system as explicit
process models, and how such models can be
exchanged, adapted or reused. As end users have
different level of technical expertise and attitudes
towards maintaining process data, we suggest that it
is important to consider possibilities for “seeding,
evolutionary growth, and reseeding (SER)” (Fisher
et al., 2004) of user-defined task structures for their
refinement and complementation.
This paper presents a distributed architecture for
supporting enterprise-wide “programming by
example” (Liebermann, 2001) of weakly-structured
process models. This EUD technique supports
unobtrusiveness by enabling generation of ad-hoc
workflows from captured, executed activities, and
yet allowing users to proactively tailor the emergent
processes at use time. The system tracks user actions
on tasks in users’ personal to-do lists and replicates
task data on central server instance. There, personal
to-do lists of different process participants are
integrated to end-to-end processes based on tracked
email exchange for task delegation. Captured data is
disseminated in different server repositories which
enables multiple perspectives on processes: process,
resource or user -centric.
3 FRAMEWORK OVERVIEW
The architecture described in this paper, supports a
framework for light-weight composition and
management of ad-hoc business processes (Stoitsev
et al., 2008). This section provides a high-level
overview of the framework to clarify the
architectural entities described further in the paper.
Aalst et al. (1999) discuss business process
flexibility by suggesting three basic dimensions of a
workflow – “case”, “process” and “resource”
dimension. Human activities thereby comprise cases,
which are handled through corresponding processes,
executed through a sequence of tasks by using
appropriate resources. Unpredictability of ad-hoc
activities implies dynamic adaptations of tasks and
resources and deviations in case handling. The used
framework considers these dimensions through its
entities: tasks, artifacts, human actors and Task
Patterns (TP) (see also Riss et al., 2005).
A framework overview is shown in Figure 1. A
case repository with reusable task structures,
generated through user activities, is shown on the
left hand side – case information is available in
individual workspaces of various users and is
additionally replicated on central enterprise
repositories as described further in this section.
Tasks A and B are part of a concrete collaborative
process. Tasks C and D represent process fragments,
elaborated by domain experts and may be part of
running processes or explicit case descriptions. On
the right hand side, the provision of resources by
external artifact managers is shown. Different type
ICEIS 2008 - International Conference on Enterprise Information Systems
166
of artifacts and their association to tasks is discussed
further in this section.
While user U
1
interacts with a hierarchical task
list (tasks A … A
2.1
depicted through gray ellipses)
in the local workspace, the system uses web services
to track and replicate all task structure and context
information on three central enterprise repositories:
(i) task tracking repository – holding task structure
and context information associated to task objects
(e.g. subject, description status etc.); (ii) artifact
repository – providing central storage for task
attachments (artifacts); (iii) user repository – storing
information about the involved persons (human
actors) that can be provided as expertise
recommendation if a task is reused later on.
Task delegation over email (A
1
Æ B) is tracked
to bind the individual task hierarchies (to-do lists) to
a Task Delegation Graph (TDG) (Stoitsev et al.,
2008), which in Figure 1 includes only two users
and is hence highly simplified. Changes on task
instances in running processes are propagated with
notifications iteratively throughout the TDG. This
allows stakeholders to adapt their corresponding
tasks according to the current situation by changing
content or status information.
Tasks may contain references to Externally
Managed Tasks (EMT) (B
1
Æ C). An EMT
represents a possible execution (breakdown) of a
given task. An EMT reference can be used to fetch
the complete EMT task hierarchy including context
information of the contained task instances from the
server. An EMT can contain further EMT
references. The reference chain ends with a task
without further references (task D). Changes on
EMT trigger notifications to the reusing task
instances. Stakeholders can update to the new
structure or preserve the locally used EMT copy and
release the reference, which corresponds to an apply
TP operation as discussed further in this section.
A task instance can contain different types of
artifacts. Artifacts are files, e.g. Microsoft
Word/Excel documents, archive (zip) files or even
executable files, which are used or generated during
task execution and are associated to tasks e.g. as
attachments. Artifacts can be: (i) locally managed,
non-externalized i.e. not replicated in a remote
repository (black circle in A
2.1
in the client
workspace) – used to store confidential resources;
(iii) Externalized Artifact (EA) (white circle in A
2.1
represents local EA reference to an EA in the artifact
repository – gray circle with a black outline). EAs
are implicitly replicated and matched based on
binary checksums on the server and allow detection
of common tasks based on the usage of similar
resources (e.g. C
2
and D use the same EA), which
assists towards document-based PID (Holz et al.,
2007); (iii) Externally Managed Artifact (EMA)
(white circle with a black dotted outline in D
2.2
represents an EMA reference to an EMA in the
artifact repository - gray circle with a black dotted
outline). EMAs are notifications-enabled and allow
C
2
C
1
P
1
A
1
P
2
D
2
D
2
B
1
B
2
D
2.2
D
1
D
1
C
1
B
2
B
1
A
1
P
1
Email
External Task
Managers
Task Requester
LOCAL
USER
WORKSPACES
TASK
REPOSITORY
ARTIFACT
REPOSITORY
Task Recipient
U
1
U
2
External Artifact
Managers
G
1
U
3
U
4
tracking
U
x
Task Pattern (XML)
Case Expert
PROCESS
CASE
REPOSITORY
(GUIDANCE)
ancestor
descendant
USER
REPOSITORY
U
1
U
2
U
3
U
4
owner /
recipient
P
P
2
P
2.1
P
P
2.1
U
x
A
A
2
A
2.1
A
A
2
A
2.1
B
B
C
C
2
C
1.1
C
C
1.1
D
D
2.1
D
2.2
D
D
2.1
ÆP
1
ÆP
1
ÆP
ÆP
Æ
Æ
U
1G1
U
nG1
author
Figure 1: Framework overview.
ARCHITECTURE FOR END USER-DRIVEN COMPOSITION OF UNDERSPECIFIED, HUMAN-CENTRIC
BUSINESS PROCESSES
167
dynamic synchronization of artifacts used in
personal tasks with changes of domain experts. As
an extension to the original framework we suggest
storing data for EMA author in the user repository.
Task-related expertise is stored through basic
user information - user name and id (email address).
Two roles are supported: (i) owner – pointing at the
person, whose individual to-do list contains the task
(e.g. U
1
); (ii) recipient(s) – pointing at the person(s),
who have received a delegated task in a
collaborative process (e.g. U
2
).
The reuse of emerging task hierarchies is
enabled through TPs. A TP contains a task with its
complete sub task hierarchy and context information
of all contained task instances, including artifacts
and human actors. A TP can be exported from an
arbitrary task in the to-do list in the local workspace
or a complete TDG under a given task can be
exported from the task tracking repository. TP can
be stored to local or remote Task Pattern
Repositories (TPR). Local TPR are XML files
(Stotisev et al., 2008). Remote TPR store global
best-practices (TPs) and are extensions to the
original framework, enabling activation of end-to-
end process examples throughout the enterprise
without exchanging TP XML files but through
referring to centrally stored TPs. One implicit
central TPR is the task tracking repository.
When a TDG is extracted, personal task
hierarchies of different process participants are
saved as separate TPs and requester tasks receive
EMT references to the corresponding recipient TPs.
The recipient tasks are generally considered as
suggested TPs for further handling of the requester
tasks. Weakly-structured process models hence
emerge as TPs (process fragments) which are
interlinked based on suggestions. The collaborative
flow is defined through suggested recipients in tasks.
Applying a TP creates the complete task
structure in the local user workspace by feeding all
context information and setting all references to
suggested TPs (EMTs) and used artifacts. EAs and
EMAs are fetched from the artifact repository and
attached to the local tasks in the workspace. All
tasks resulting form TP application are accordingly
replicated on the task tracking repository. Users can
change the prescribed collaborative flow of an
enacted process example by entering different than
the suggested recipients. References to suggested
TPs (EMTs) can be forwarded with delegations or
used by the task owner to fetch the corresponding
EMT and execute the tasks themselves. This allows
opportunity-based and emergent adaptations of
enacted process examples. If a user applies a TP
with suggested TP references (EMTs) from a local
TPR, the locally referenced structures can be
exported iteratively in a default, user-specific remote
TPR so that they can be reused by other users.
To allow tracing of deviating cases,
ancestor/descendant relationships are maintained.
These are set iteratively in task hierarchies, resulting
from TP application (ÆP in A, ÆP
1
in A
1
etc.).
Copy/paste operations of task fragments executed
during TP editing in a visual environment also
produce such references. These references allow
comparison of resulting with originating entities
(ancestor – descendant) and comparison of resulting
entities with same origin (descendant – descendant).
As each task instance preserves its ancestor
reference, fine-grained, instance-based traceability
of task reuse is enabled.
4 ARCHITECTURE
This section presents the architecture supporting the
above framework. The term “office applications”
used in the following is a conceptual term and does
not explicitly refer to Microsoft Office or imply the
features offered by this environment. The presented
architecture is implemented in the CTM prototype,
which is integrated in Microsoft Outlook (OL) by
exploiting the fact that tasks and email are offered in
the same office application.
The system has a three-tier architecture
consisting of client, server and persistence layers as
shown in Figure 2. The client layer contains the
logic for the personal task management. The server
layer comprises the components, providing the
tracking functionality, the overall repository access
and data retrieval, and the business logic (e.g.
notifications handling). The persistence layer
encompasses the runtime data storage for task
tracking and the user, artifact and TP repositories.
4.1 Client Layer
The client-side components are integrated in the
office applications environment of the end user. The
focus is set on the email client, which may contain
also certain task management and calendar
functionality.
4.1.1 Office Applications Integration Layer
The task management system is coupled to the office
applications over an Office Applications Integration
Layer (OAIL). The latter enables the usage of email
ICEIS 2008 - International Conference on Enterprise Information Systems
168
for the purposes of the task management system, by
serving as a proxy and enabling email pre-formatting
of outgoing emails and appropriate handling of
incoming emails with task-related content. If task
management functionality is included to some extent
in the office applications environment, the
integration layer should preferably make use of it by
enabling tracking of task-related operations (e.g.
edit, create, delete) through web services. The
integration layer should also provide functionality to
embed the user interface of the task management
system in the office applications environment.
Concretely the CTM front-end is delivered as an OL
Add-In, where the OAIL comprises classes with
proprietary extensions of OL mail and task items,
adding custom properties and a set of event handlers.
Office Applications
Tracking
Repository
TaskMail Calendar
Office Applications Integration Layer
Plug
WS Clients
Mail Server
WS Clients
External
Web
Services
Task
Management
Web Client
Local TaskPattern
Repositories
Persistence
Layer
Server
Layer
Web Services
Client
Layer
Task Pattern
Repositories
Task Management Server
Task Pattern Service
Repository Managers
Task Patterns
Explorer/Editor
Office-Integrated Task
Management Client
Artifact
Repositories
Artifact Service
Repository Managers
Local Artifact
Repositories
HTTPSMTP/IMAP
SMTP/IMAP
SQL SQL/File I/O
SOAP/HTTP
Tracking Service
SQL/File I/O
Artifacts
Explorer
User
Repository
SQL
Figure 2: System architecture.
4.1.2 Office-Integrated Task Management
Client (OITMC)
The Office-Integrated Task Management Client
(OITMC) holds the complete presentation logic for
the task management system within the selected
office application (i.e. the email client). The OITMC
also contains the composition environment for the
personal to-do lists. Additional extended
functionality for the inspection and adaptation of
existing task structures is provided through a Task
Pattern Explorer/Editor (TPE). It enables editing of
local TPs (exported to XML files) as well as the
complete functionality for creating, retrieving and
editing remote TPRs and TP instances. Artifact
handling is provided in an Artifact Explorer (AE)
component. It enables users to add EMAs and
explore their version and history and to further
explore EAs and their referencing tasks through
querying data from the server.
Important parts of the OITMC are the task
management Web Service (WS) clients. These are
responsible for tracking of task related actions on the
task management server and executing updates and
queries on the remote repositories. Additional WS
clients can be plugged in the OITMC to execute
task-related operations on external systems e.g. to
trigger transactions in a workflow or Enterprise
Resource Planning (ERP) system. A simple
approach for binding such external WS in tasks is
e.g. through executable attachments (artifacts).
4.1.3 Local Task Pattern Repositories
Although reuse of task structures should be utilized
on enterprise basis, some task structures can still be
stored locally. On the one hand local TPs can hold
personal best-practices in a private, non-distributed
manner. On the other hand, a local TPR may be used
by caching mechanisms, included in the OITMC, to
store local copies of once retrieved remote TPs in
order to avoid communication overhead for repeated
TP retrieval. Currently local TPRs are XML files,
containing one or more TP definitions as described
by Stoitsev et al. (2008) in an additional repository
element.
4.1.4 Local Artifact Repositories
Artifacts are generally managed in globally
accessible, remote Artifact Repositories (ARs). At
the same time, artifacts can be stored also within
local AR to ensure that sensitive documents will not
be implicitly externalized throughout the enterprise -
such are the locally managed, non-externalized
artifacts. Local ARs can be organized as file system
folders. Additionally, the OITMC can include
caching mechanisms for storing once retrieved
remote artifacts in a local AR, from where these will
be retrieved on subsequent requests.
4.1.5 Task Management Web Client
(TMWC)
The TMWC provides overview of the evolving task
structures beyond the boundaries of the personal
workspace. It allows inspection of the complete
TDGs, where all attributes of the task nodes can be
viewed – status, percent complete, due date etc. as
ARCHITECTURE FOR END USER-DRIVEN COMPOSITION OF UNDERSPECIFIED, HUMAN-CENTRIC
BUSINESS PROCESSES
169
well as artifact lists (attachments). Detailed
overview of the dialog flow for task delegation is
enabled in the TMWC. The overall purpose of this
component is to provide system independent
environment for process navigation, where advanced
users with higher process expertise can inspect work
distribution and identify potential bottlenecks and
optimization possibilities. The TMWC in CTM can
be shown through “Process Info” links in tasks and
emails in the OITMC. Although in CTM the TMWC
currently provides only an overview of the emerging
processes, in advanced implementations this
component may include functionality for dynamic
task changes in the displayed, shared-accessible task
hierarchies (cf. Bernstein, 2000). These should be
accordingly reflected in the OITMC and supplied
with notifications.
4.2 Server Layer
The server layer encompasses two major
components – the mail server and the task
management server. External WS e.g. from ERP
systems (cf. 4.1.2) can also be considered in this
layer. The different components from the server
layer do not necessarily reside on the same host.
4.2.1 Mail Server
The mail server is implicitly included in the
architecture as the exchange of tasks is realized over
email, i.e. this component is used by the office
applications email client and could be e.g. a
Microsoft Exchange server. Furthermore, email can
be used for the propagation of notifications from the
back-end to the clients. Therefore a bidirectional
relation from the task management server to the mail
server is depicted in Figure 2. However the
presented architecture does not rely exclusively on
“computational email” as known email-based
workflows (Agostini et al., 1997) but utilizes WS to
increase performance and extensibility.
4.2.2 Task Management Server
The task management server application provides all
basic services for the system and realizes the
connection of the clients to the remote repositories.
The Tracking Service updates the persistent state
of tasks on the server when these are created or
updated on the client. The service handles
additionally the collaborative flow. All task-related
email exchange is stored in dialog instances, mapped
to the appropriate requester and recipient tasks. All
messages are available in the tracking repository
with text and attachments. The attachments are
replicated to a remote AR. The tracking service
feeds also owner and recipient data in the user
repository to dynamically fill the expertise database.
Notifications on task instance changes (cancel,
complete, delete request) are also propagated though
the tracking service to allow stakeholder adaptations.
The support of offline system usage (personal task
adaptations in a disconnected client) is a complex
issue and is not discussed explicitly in the presented
paper. This mode is generally supported through
tracking buffers on client and server side. On client
side such is a local tracking record file. On server
side, unconsumed notification events are kept in a
DB table. Switching in online mode triggers
evaluations of the client and server records, aiming
to avoid collisions on task actualization e.g. increase
of percent complete for a delegated task cancelled
by the requester. The compensation mechanisms and
collision prevention are not final and still under
research. However, synchronous tracking support
seems currently sufficient as CTM test use takes
place in offices, where users are online.
The Task Pattern Service updates remote TPs,
enables search in the remote TPRs and delivers
remote TP structures to the clients. It further
provides functionality for delivering notifications on
reused, suggested TPs (EMTs). Multiple Repository
Managers can realize the connection to multiple
TPRs of the same or different kind (e.g. database or
file system based). User data is fed to the user
repository analogously to the task tracking.
The Artifact Service provides functionality for
the replication of EAs and for the management of
EMAs on remote ARs. This includes all update,
search and retrieval functionalities. The service can
handle different repository types (e.g. database or
file system based) through appropriate artifact
Repository Managers. Author information of EMAs
is stored in the user repository.
The Tracking Service and the TP Service
communicate with each other to enable setting and
retrieval of ancestor/descendant relationships
between tracked tasks and explicit TP instances.
Both services communicate with the Artifact
Service, to maintain associations of artifacts within
active (tracked) tasks and TPs. All services have
access to the user repository to feed expertise (task
owner/recipient) and (EMA) author information. In
CTM all services are based on the Simple Object
Access Protocol (SOAP) and integrated into a java
application server.
ICEIS 2008 - International Conference on Enterprise Information Systems
170
4.3 Persistence Layer
The persistence layer comprises the task tracking
repository and the user, artifact and TP repositories.
Components used to store different types of data can
reside on physically different hosts. This would
require additional repository mapping functionality
in the respective services for relating the different
entities – tasks, users, artifacts and TP. Currently
CTM implements the different repositories as tables
in a single DB for simplicity reasons. Thereby
entities from the different repositories are associated
through foreign-key relations.
4.3.1 Tracking Repository
This repository stores the data, generated through
tracking of client-side user operations, on the server.
The latest state of all user tasks from the personal
workspaces is replicated in this repository. The
tracking repository provides the input for the web-
based overview of a TDG (generated enterprise
process flow) in the TMWC and stores also all task-
related dialogs, generated through task delegation.
4.3.2 User Repository
The user repository stores human actor information.
Such can be added in the following ways: (i) through
tracking of evolving tasks in running processes i.e.
task creation feeds owner data, task delegation
delivers recipient information; (ii) saving of a
remote TP, containing human actor information
(owner, suggested recipients); (iii) adding and
editing of EMA adds author information. This
approach enables user-based navigation throughout
document and task repositories and tracing what
process fragments and resources are accessed or
submitted by a given user. No domain-specific roles
are introduced currently as this could harm the
generic character of the approach and its ability to
support various user activities in different
enterprises from different business domains.
4.3.3 Remote Task Pattern Repositories
A remote TPR holds reusable task structures. More
than one repository can be maintained within the
system. One implicit TPR is the tracking repository.
However, while task structures in the tracking
repository may be changed frequently when a user
updates their personal to-do list in the OITMC, the
idea of a remote, central TPR is to keep reusable
process fragments in a more consistent manner and
to provide global best-practices. As a TP can
represent an EMT, TP changes must be supported
with appropriate notifications handling to allow
stakeholder adaptations. Remote TPRs can be
implemented using DB tables. Access and
management of different repository types should be
provided in a unified manner by the Task Pattern
Service. Currently CTM supports only one remote,
DB-based TPR in parallel to the tracking repository.
4.3.4 Remote Artifact Repositories
A remote AR holds globally reusable artifacts. One
or more ARs of the same or different kind, e.g. DB
or file system based, should be managed in a unified
manner through the Artifact Service. Currently CTM
uses a single combined AR – a DB table with
artifact context information as name, checksum,
version etc. and links to actual artifact content (files)
on the server file system. All remote artifacts – EAs
and EMAs are currently kept in the same repository
(DB table) where EMAs have additional version
attributes. This can enable conversion of EA to
EMA through extending the EA attributes’ set.
5 CONCLUSIONS & FUTURE
WORK
The paper presents an architecture which
implements an integrated approach, leveraging user
experience with standard office tools for
collaboration and task management towards the
definition of weakly-structured process models by
end users. The underlying approach aims at ensuring
a “gentle slope of complexity” for users engaging in
process tailoring activities. To achieve that, the
architecture enables enterprise-wide “programming
by example” of distributed-emergent, ad-hoc
processes through web service-based activity
tracking. An important aspect is the provided
possibility for “seeding, evolutionary growth, and
reseeding (SER)” of weakly-structured process
models in shared enterprise repositories. This
enables refinement of generic processes towards
global best-practice definitions in organizations. The
presented architecture enables SER by distributing
and interconnecting data in various repositories –
tracking, user, artifact and TP repositories. This
enables different perspectives on processes: (i)
process perspective – describing the end-to-end
process flow where tasks contain all relevant artifact
and human actors’ information and additionally
allow tracing of evolutionary task reuse
(ancestor/descendant hierarchies); (ii) resource
ARCHITECTURE FOR END USER-DRIVEN COMPOSITION OF UNDERSPECIFIED, HUMAN-CENTRIC
BUSINESS PROCESSES
171
perspective - enabling detection of similar tasks
based on similar resources, i.e. exploration of EA
and EMA references in tasks; (iii) human actor
perspective – enabling tracing of user expertise and
contributions based on associated tasks and
resources, i.e. task owner/recipient, EMA author
relationships. These extensive data mining
capabilities and the enabling of end users to model
and enact end-to-end process execution examples by
combining existent and new TP, user and artifact
data go beyond known email-based and evolutionary
workflows.
A long term CTM evaluation is planned which
will allow scalability assessments for the entities,
managed through this architecture. This evaluation
should further reveal possibilities for automatic
detection of rigidly recurring process facets. We
further plan to investigate the mapping of user-
defined process descriptions to formal process
modeling notations towards automation of rigidly
recurring process fragments.
ACKNOWLEDGEMENTS
The work, this paper is based on, was supported
financially by the German Federal Ministry of
Education and Research (project EUDISMES,
number 01 IS E03 C).
REFERENCES
Aalst, W.M.P. v. d., Basten, T., Verbeek, H. M. W.,
Verkoulen, P. A. C., Verhoeve, M., 1999. Adaptive
Workflow: On the interplay between flexibility and
support. In ICEIS’99, Setubal, Portugal, pp. 353-360.
Agostini, A., De Michelis, G., 1997. Rethinking CSCW
systems: the architecture of Milano. In ECSCW’97,
Springer, pp. 33-48
Bellotti, V., Dalal, B., Good, N., Flynn, P., Bobrow, D. G.,
Ducheneaut, N., 2004. What a To-Do: Studies of Task
Management towards the Design of a Personal Task
List Manager. In CHI’04, ACM Press, pp. 735-742.
Bellotti, V., Ducheneaut, N., Howard, M., Smith, I.,
Grinter, R., 2005. Quality Versus Quantity: E-Mail-
Centric Task Management and Its Relation With
Overload. Human-Computer Interaction, Lawrence
Erlbaum Associates, Vol. 20 (2005), pp. 89-138.
Bernstein, A., 2000. How Can Cooperative Work Tools
Support Dynamic Group Processes? Bridging the
Specificity Frontier. In CSCW’00, ACM Press, pp.
279-288.
Fischer, G., Giaccardi, E., Ye, Y., Sutcliffe, A.,
Mehanjiev, N., 2004. Meta-Design: A Manifesto for
End-User Development. Communication of the ACM,
Vol. 47, No. 9 (Sep. 2004).
Gartner Research, 2006. Person-to-Process Interaction
Emerges as the 'Process of Me’, Gartner Inc.
Holz, H., Maus, H., Bernardi, A., Rostanin, O., 2005.
From Lightweight, Proactive Information Delivery to
Business Process-Oriented Knowledge Management.
Journal of Universal Knowledge Management, Vol. 0,
No. 2 (Nov. 2005), pp. 101-127.
Jorgensen, H. D., 2004. Interactive Process Models. Ph.D.
Thesis, Norwegian University of Science and
Technology, Trondheim, Norway.
Lieberman, H., 2001. Your Wish is My Command:
Programming by Example, Morgan Kaufmann.
Lieberman, H., Paterno, F., Wulf, V., 2006. End-User
Development, Springer.
MacLean, A., Carter, K., Lövstrand, L., Moran, T, 1990.
User-tailorable systems: pressing the issues with
buttons. In CHI90, ACM Press, pp. 175-182.
Riss, U., Rickayzen, A., Maus, H., van der Aalst, W.M.P.,
2005. Challenges for Business Process and Task
Management. Journal of Universal Knowledge
Management, Vol. 0, No. 2 (Nov. 2005), pp. 77-100.
Schwarz, S., Abecker, A., Maus, H., Sintek, M., 2001.
Anforderungen an die Workflow-Unterstützung für
wissensintensive Geschäftsprozesse. In WM’01, 1st
Conference for Professional Knowledge Management,
Baden-Baden, Germany.
Stoitsev, T., Scheidl, S., Spahn, M, 2008. A Framework
for Light-Weight Composition and Management of
Ad-Hoc Business Processes, LNCS 4849, Springer, pp.
213–226.
Wiig, K. M., 2004. People-focused knowledge
management: how effective decision making leads to
corporate success, Elsevier Butterworth–Heinemann.
Wulf, V., Jarke, M, 2004. The Economics of End-User
Development. Communications of the ACM, Vol. 47,
N. 9 (Sep. 2004).
ICEIS 2008 - International Conference on Enterprise Information Systems
172