Demonstrating Flexible Support for
Knowledge-Intensive Processes with proCollab
Nicolas Mundbrod and Manfred Reichert
Institute of Databases and Information Systems
Ulm University, Germany
{nicolas.mundbrod,manfred.reichert}@uni-ulm.de
http://www.uni-ulm.de/dbis
Abstract.
Knowledge-intensive processes (KiPs) are driven by knowl-
edge workers utilizing their skills, experiences, and expertise. As these
processes are emergent and unpredictable, their support constitutes a
big challenge. For coordinating and synchronizing the various tasks of
a KiPs, knowledge workers still rely on simple task lists like, e.g., to-do
list or checklists. Though task lists are intuitive, their current implemen-
tations are very ineffective: tasks are neither made explicit nor are they
personalized or synchronized. In addition, no task management lifecycle
support is provided and media disruptions frequently occur. This tool
demonstration presents the proCollab framework supporting a wide range
of KiPs (e.g., projects and cases) through integrated, lifecycle-based task
management. In particular, proCollab task trees support the provision of
task list templates and to constitute digital task lists of any kind. For
customization, it further allows integrating domain-specific methodologies
as well as configuring task lists at design and run time. Overall, the pro-
Collab framework shall improve coordination among knowledge workers,
increase work awareness, and record valuable coordination efforts.
Keywords:
knowledge-intensive processes, task management, task list,
checklist, to-do list, knowledge worker
1Introduction
Residing in sensitive key business areas, such as research, healthcare, or engi-
neering, knowledge-intensive processes (KiPs) have become the centerpiece for
creating value in many companies [3]. While driving KiPs, knowledge workers uti-
lize their skills, experiences, and expertise to cope with sophisticated tasks. Thus,
the systematic and sustainable support of KiPs constitutes a prerequisite for
achieving business goals. KiP support, however, still poses a big challenge. KiPs
can be characterized as non-predictable,emergent,goal-oriented, and knowledge-
creating processes [1,3]. Consequently, KiPs have not been fully supported by
contemporary process-aware information systems so far. Instead, knowledge
workers still rely on simple task lists (e.g., to-do lists, checklists) to coordinate
their work [6]. But these instruments are error-prone and ineffective. Moreover,
2Nicolas Mundbrod and Manfred Reichert
knowledge workers suffer from media disruptions and a lack of lifecycle-based
task management preventing them from reusing existing artifacts (e.g., task lists)
in similar KiPs [4].
The proCollab
1
framework aims at the systematic and sustainable support of
KiPs. As tasks constitute the key objects regarding planning and quality assurance
for knowledge workers, proCollab provides process- and lifecycle-based task
management [3], empowering knowledge workers to collaborate more effectively.
To make use of best practices and knowledge gained in the context of comparable
KiPs, proCollab provides process and task list templates, which knowledge workers
may instantiate on demand. To foster the reuse of these templates, a context-
aware approach for configuring task list templates is included as well. Knowledge
workers may then easily configure task lists either at design or run time. To
increase work awareness in addition, proCollab encompasses a flexible state
management concept. The latter may be customized to optimally align KiPs,
task lists, and tasks with domain-specific requirements. Finally, the proCollab
framework aims to improve coordination among knowledge workers, prevent
media disruptions, and record valuable coordination efforts and knowledge.
2proCollab and its Significance to the BPM field
To enable a generic, but customizable support for KiPs, the proCollab framework
relies on processes,task trees, and tasks as the conceptual pillars for representing a
wide range of KiPs as well as task-based artifacts used by knowledge workers. To
establish and provide lifecycle-based task management support in the context of
KiPs [2], processes are refined to process templates and process instances, whereas
task trees are refined to task tree templates (with task templates) and task tree
instances (with task instances) respectively. Figure 1illustrates the relationship
and interplay of those components.
Task Tree
Templates
with
Task Templates
0-n
0-1
0-1
0-n
0-n
1-n 1-n
1-n 1-n
0-n 0-n
0-n 0-n
0-n 0-n
0-n
Subordinated
Process Templates
Subordinated
Process Instances
Process
Templates
B
Sub-Task Tree
Templates with
Task Templates
2
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Task Tree
Root
1 B
Root
A B
A1 A2 B1 B2
Task Tree
Task Tree
Root
A B
A1 A2 Root
B1 B2
Task Tree Instances
with Task Instances
Process
Instances
Sub-Task Tree
Instances with
Task Instances
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Task Tree
Root
B1B2
Root
A B
A1 A2 B1 B2
Task Tree
Task Tree
Root
A B
A1 A2 Root
B1 B2
Instantiation
Optimization
Process
Plan
Do
Study
Act
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Process
Plan
Do
Study
Act
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Process
Plan
Do
Study
Act
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Process
Plan
Do
Study
Act
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Process
Plan
Do
Study
Act
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Process
Plan
Do
Study
Act
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Root
A B
A1 A2 B1 B2
Task Tree
Fig. 1.Overview of proCollab Components
Process templates and task tree templates enable knowledge workers to ac-
celerate planning of their tasks based on best practices. Every process template
may have several subordinated process templates and feature various properties,
1Process-aware Support for Collaborative Knowledge Workers
The proCollab Demonstrator 3
conditions, and linked resources. Every process template may be linked to an
arbitrary number of task tree templates. A task tree template, in turn, contains
task templates and, optionally, subordinated task tree templates. Thereby, it
reflects best practices for planning (to-do list) or quality assurance (checklist)
in the context of KiPs. For example, process template for clinical surgeries may
contain a task tree template for ensuring patient safety during surgeries.
At run time, knowledge workers collaborate in the context of process instances.
A process instance may represent a running project, a case, or any kind of knowl-
edge worker collaboration. Moreover, it has properties like a goals and it features
resources (e.g., documents). A process instance may have subordinated process
instances to let knowledge workers focus on specialized sub-goals. Every process
instance may comprise multiple task tree instances with corresponding task in-
stances. A task tree instance constitutes the generic representation of common
task-based artifacts (e.g., a to-do list) in use. For example, an automotive engi-
neering project with to-do lists for planning and checklists for quality assurance
can be supported by a respective proCollab process instance with corresponding
task tree instances (of type “to-do list” and “checklist”) (cf. Figure 2).
Name: E/E Car Component Development Project
2 Years A�er Start
...
Due Date:
Goal: E/E Car Component Sa�sfying Specifica�on
Subordinated Process Templates
Requirements
Engineering
System
Design ...
Linked Resources
Specifica�on
Templates
...
Design
Guidelines
Quality Assurance
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist Templates
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Planning
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
To-do List Templates
To-do A
To-do A1
To-do A2
To-do B
To-do B1
To-do B2
Subordinated Process Templates
Requirements
Engineering
System
Design ...
Linked Resources
...
Specifica�ons Designs
Status: Running
Name: Development Project for Airbag Control Unit ACU-5891 V14
15 October 2016
...
Due Date:
Goal: ACU-5891 V14 Sa�sfying Specifica�on CCS-8641
Quality Assurance
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist Instances
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Planning
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
To-do List Instances
To-do A
To-do A1
To-do A2
To-do B
To-do B1
To-do B2
Subordinated Process Templates
Requirements
Engineering
System
Design ...
Linked Resources
...
Specifica�ons Designs
Status: Running
Name: Development Project for Airbag Control Unit ACU-5891 V14
15 October 2016
...
Due Date:
Goal: ACU-5891 V14 Sa�sfying Specifica�on CCS-8641
Quality Assurance
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist Instances
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Planning
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
To-do List Instances
To-do A
To-do A1
To-do A2
To-do B
To-do B1
To-do B2
Subordinated Process Templates
Requirements
Engineering
System
Design ...
Linked Resources
...
Specifica�ons Designs
Status: Running
Name: Development Project for Airbag Control Unit ACU-5891 V14
15 October 2016
...
Due Date:
Goal: ACU-5891 V14 Sa�sfying Specifica�on CCS-8641
Quality Assurance
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist Instances
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Planning
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
To-do List Instances
To-do A
To-do A1
To-do A2
To-do B
To-do B1
To-do B2
Subordinated Process Templates
Requirements
Engineering
System
Design ...
Linked Resources
...
Specifica�ons Designs
Status: Running
Name: Development Project for Airbag Control Unit ACU-5891 V14
15 October 2016
...
Due Date:
Goal: ACU-5891 V14 Sa�sfying Specifica�on CCS-8641
Quality Assurance
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Checklist Instances
Check Item A
Check Item A1
Check Item A2
Check Item B
Check Item B1
Check Item B2
Planning
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
Task List
Task A
Task A1
Task A2
Task B
Task B1
Task B2
To-do List Instances
To-do A
To-do A1
To-do A2
To-do B
To-do B1
To-do B2
a) Process Templates of Specializa�on Type Project Instantiation b) Process Instances of Specializa�on Type Project
Fig. 2.
Visualization of Process Templates and Instances from the Automotive Domain
In general, knowledge workers may create a process instance based on a
process template or starting without any template at all. If a process template is
instantiated, the linked task tree templates are instantiated, too. Further, the
created task tree instances are linked to the corresponding process instance. In
the context of a particular process instance, knowledge workers may instantiate
further task tree templates or add “blank” task tree instances on demand.
To support a wide range of application scenarios, proCollab employs spe-
cialization types enhancing the generic data structures of processes and task
trees. Depending on the chosen specializations, processes (e.g., projects) and task
trees (e.g., to-do lists) may feature additional properties (e.g., states), conditions,
constraints, or assignments. For example, a proCollab process may be adapted to
a specific automotive project (cf. Fig. 2). If a task tree instance is linked to the
specialization type to-do list, it will be interpreted as a to-do list instance with
corresponding user interface representations and properties. To ensure a coherent
use of specialization types, the latter may be linked to each other. For example,
the specialization types to-do list and to-do item are interlinked and, hence, task
trees of type to-do list may only contain tasks of type to-do item. To further
4Nicolas Mundbrod and Manfred Reichert
increase domain-specific support, proCollab employs a generic state management
for its stateful key components. This enables us to integrate domain-specific
methodologies as well as to manage different types of proCollab components in
a controlled manner. Additionally, proCollab provides a configuration concept
for task tree templates. The latter enables the efficient configuration of task list
templates in accordance to the given application context. Due to a lack of space,
both the state management and the configuration concept are presented in the
screencast of this demonstration in more detail (cf. Section 3).
Concerning the significance of proCollab to the BPM field, the design of a
systematic KiP support still constitutes a big challenge. While predictable business
processes can be well supported by process-aware information systems based on
pre-specified process models (doing-by-design), unpredictable and emergent KiPs
require a degree of flexibility traditional systems are not able to provide due to the
limitations of model-driven approaches. In turn, the proCollab framework relies
on the design-by-doing approach necessitating different concepts and functions
to support a wide-range of KiPs in a generic, but still domain-specific way.
In comparison to declarative process execution systems like Declare or DCR
Graphs [7] and approaches based on CMMN [5], proCollab focuses on the stateful
and flexible task (list) management support for KiPs and knowledge workers. In
particular, knowledge workers themselves shall drive proCollab processes and, as
a consequence, the management of tasks and task lists, too.
3The proCollab Tool and its Maturity
To prepare empirical studies and to evaluate the technical feasibility of the
proCollab framework, we developed the proCollab tool. The latter is realized
with Java EE 7and relies on a MVC-based architecture (cf. Fig. 3). In particular,
the application logic layer represents the core of the tool realizing the key services
of the proCollab framework. The REST-based interface enables web and mobile
applications to communicate with these key services. Finally, the web application
based on AngularJS provides state-of-the art user interfaces (cf. Fig. 3) to the
knowledge workers interacting with the proCollab tool.
The current version of the proCollab tool enables knowledge workers to
manage the proCollab key components, i.e., they may manage KiPs in the
shape of common projects or cases (i.e., proCollab processes) including task
trees embodied as, e.g., to-do lists or checklists. Moreover, knowledge workers
may configure task tree templates or instantiate process/task tree templates
on demand (see screencast). Furthermore, the proCollab tool allows integrating
domain-specific state models for templates and instances of processes, task trees,
and tasks. In turn, the assignment of users to tasks, the integration of process
resources (e.g., documents, structured data), and the synchronization of tasks and
processes are currently in development to be integrated in the web application.
Based on its architecture and the technologies in use, the proCollab tool can scale
up well and even be deployed in cloud-based environments to serve a significant
amount of knowledge workers concurrently.
The proCollab Demonstrator 5
a) b)
Mobile Applica�onsWeb Applica�on
REST API
Task Tree Services
(Instances)
Presenta�on
Database Management Systems
Java Persistence API
Process Services
(Instances)
Process Template
Repository
Task Tree Template
Repository
noitacinummoC
ecnetsisreP
User and Role
Management
Data Management
Applica�on
Java Content Repository
Fig. 3.Architecture and Screenshot of the proCollab Tool
Finally, we prepared a screencast available on
bpm2017demo.procollab.de
to
demonstrate proCollab. The screencast illustrates the key aspects of the proCollab
framework using the scenarios of developing a website and conducting a surgery.
4Conclusion
This demo presented the proCollab framework and its proof-of-concept tool. The
latter is implemented as a scalable, cloud-ready, and web-based application plat-
form. This allows users to collaborate in the scope of KiPs and to simultaneously
manage proCollab key components to receive optimal KiP support. Based on the
tool, user experiments will be conducted to systematically analyze whether KiPs
can be successfully evolved when using the proCollab framework. Overall, we
believe that the proCollab framework will improve coordination among knowledge
workers, increase work awareness, and record valuable coordination efforts.
References
1.
Di Ciccio, C., Marrella, A., Russo, A.: Knowledge-Intensive Processes: Characteristics,
Requirements and Analysis of Contemporary Approaches. J on Data Semantics 4(1),
29–57 (2014)
2.
Mundbrod, N., Beuter, F., Reichert, M.: Supporting Knowledge-Intensive Processes
through Integrated Task Lifecycle Support. In: Proc. EDOC 2015. pp. 19–28 (2015)
3.
Mundbrod, N., Kolb, J., Reichert, M.: Towards a System Support of Collaborative
Knowledge Work. In: BPM 2012 Workshops. LNBIP 132 (2013)
4.
Mundbrod, N., Reichert, M.: Process-Aware Task Management Support for
Knowledge-Intensive Business Processes: Findings, Challenges, Requirements. In:
Proc. EDOCW’14. pp. 116–125 (2014)
5.
Object Management Group: Case management modeling and notation (cmmn) 1.1
(2016), http://www.omg.org/spec/CMMN/1.1/
6.
Pryss, R., Mundbrod, N., Langer, D., Reichert, M.: Supporting medical ward rounds
through mobile task and process management. Inf Sys and e-Business Management
13(1), 107–146 (2015)
7.
Reijers, H.A., Slaats, T., Stahl, C.: Declarative Modeling—An Academic Dream or
the Future for BPM? In: Proc. BPM 2013, pp. 307–322. Springer (2013)
