presented by Barbara Weber Parts of this research was funded by the Austrian Science Fund (FWF): P23699-N23.
Overcoming the challenges of reuse in
large process model repositories
03.09.2012 rBPM 2012, Tallinn 2
Agenda
Motivation Process of Process Modeling Supporting the Process of Process Modeling Refactoring Test driven Modeling Literate Process Modeling
Summary
MOTIVATION quality of process models as a necessary precondition for reuse
03.09.2012 rBPM 2012, Tallinn 4
Reuse of Common Process Knowledge
Increasing adoption of process-aware information systems
Emergence of large process model repositories often comprising a large number of related process fragments
Reuse of common process knowledge to reduce process modeling and maintenance efforts
03.09.2012 rBPM 2012, Tallinn 5
Quality Problems
Error rates between 10% and 50% in industrial process model collections (Mendling 2009, Mendling et al. 2008)
Impedes comprehensibility and maintainability of process models (Mendling 2008, Weber & Reichert 2008, Weber et al. 2011)
o Non intention-revealing or inconsistent naming (Mendling et al. 2010)
o Redundant process fragments (Hallerbach et al. 2010)
o Large and unnecessarily complex process models (Soto et al. 2008)
03.09.2012 rBPM 2012, Tallinn 6
Quality Problems
Error rates between 10% and 50% in industrial process model collections (Mendling 2009, Mendling et al. 2008)
Impedes comprehensibility and maintainability of process models (Mendling 2008, Weber & Reichert 2008, Weber et al. 2011)
o Non intention-revealing or inconsistent naming (Mendling et al. 2010)
o Redundant process fragments (Hallerbach et al. 2010)
o Large and unnecessarily complex process models (Soto et al. 2008)
THE PROCESS OF PROCESS MODELING
how does the process of creating, modifying, and reusing process models look like
Process of Process Modeling (PPM) • Iterative, highly flexible process • Depends on individual modeler • 3 successive phases
03.09.2012 rBPM 2012, Tallinn 8
Comprehension
Modeling Reconciliation (Pinggera et al. 2010, Pinggera et al. 2011, Pinggera et al. 2012)
• Understand requirements • Understand existing process model • Understand components for reuse • Chunking (Cant et al. 1995)
– Understanding in chunks (group of information)
Process of Process Modeling Comprehension
03.09.2012 rBPM 2012, Tallinn 9
Comprehension
Modeling Reconciliation
• Central Concept: Working Memory • Required by all conscious mental activities • Severely limited: 7 +/- 2 information „slots“ (Miller 1956) • Mental effort: utilization of working memory (Paas et al.
2003) • Overflow: rapid performance decrease! (Sweller 1988)
Process of Process Modeling Comprehension
03.09.2012 rBPM 2012, Tallinn 10
Comprehension
Modeling Reconciliation
03.09.2012 rBPM 2012, Tallinn 11
• Abstraction o hiding of irrelevant information (Parnas 1972) o supports human mind’s attention management (Larking
and Simon 1987) • Split-attention effect (Sweller and Chandler 1994)
o occurs when information from several sources needs to be integrated
o switching attention between models
Process of Process Modeling Comprehension
Comprehension
Modeling Reconciliation
03.09.2012 rBPM 2012, Tallinn 12
• External memory o mechanism for reducing mental effort o Information storage outside the human cognitive system
(e.g., pencil and paper or a blackboard)
• Cognitive Trace o Information taken from working memory and stored in an
external memory (e.g., to mark, update, and highlight information)
Process of Process Modeling Comprehension
Comprehension
Modeling Reconciliation
03.09.2012 rBPM 2012, Tallinn 13
• comprehended chunks are formalized in process model – by creating new model elements – by integrating reusable components
• varying number of modeling steps
Process of Process Modeling Modeling
Comprehension
Modeling Reconciliation
03.09.2012 rBPM 2012, Tallinn 14
• improve understandability – reorganize model (refactor) – utilize secondary notation, typographic cues
• facilitate next comprehension phase
Process of Process Modeling Reconciliation
Comprehension
Modeling Reconciliation
REFACTORING fostering reusability of large process model repositories
03.09.2012 rBPM 2012, Tallinn 16
• Improving model quality without changing the observable behavior of the model
Process Model Refactoring Reconciliation Support Fostering Reuse
Comprehension
Modeling Reconciliation
(Weber and Reichert 2008, Weber et al. 2011) 03.09.2012 17
Typical Process Model Smells PMS1: Non-intention Revealing Naming of Activities / Process Model
PMS2: Contrived Complexity
PMS3: Redundant Process Fragment
PMS4: Large Process Models
PMS5: Lazy Process Models
PMS6: Unused Branches
PMS7: Frequently Occurring Instance Changes
PMS8: Frequently Occurring Variant Changes
03.09.2012 rBPM 2012, Tallinn 18
Identification of Process Model Smells
Application of Refactoring Techniques
RF1: Rename Activity
RF2: Rename Process Schema
RF3: Substitute Process Fragment
RF4: Extract Process Fragment
RF5: Replace Process Fragment by Reference
RF6: Inline Process Fragment
RF7: Re-label Collection
RF8: Remove Redundancies
RF9: Generalize Variant Change
RF10: Remove Unused Branches
RF11: Pull Up Instance Change 19
Process Model Refactorings
(Weber and Reichert 2008, Weber et al. 2011)
03.09.2012 rBPM 2012, Tallinn 20
Labeling of Process Models
• PMS1. Non intention revealing naming of activities / process models
o Ambiguous or non intention revealing labels
o Inconsistent use of labeling styles
(Weber et al. 2011)
03.09.2012 rBPM 2012, Tallinn 21
Number of ways students name an activity in a process model • Insights from a process modeling experiment
with 113 students.
• The following sentence in the process description resulted into 84 different ways for naming this particular activity:
“Afterwards the scouting team attends games of the player they are interested in live in the football stadium.”
(Fahland 2012)
03.09.2012 rBPM 2012, Tallinn 22
Inconsistent Names and Labeling Styles • Repository with 70 process models from healthcare • 16 out of 70 process models contained activities
regarding the scheduling of medical procedures (e.g., surgeries, medical examinations, drug administration)
• Although activities had similar intentions, different labels and labeling styles were used o “Make appointment”, “appointment”, “schedule examination”,
“fix day”, “agree on surgery date”, “plan”
(Weber et al. 2011)
03.09.2012 rBPM 2012, Tallinn 23
Labeling Conventions + Refactoring of Labels • Usage of labeling conventions and domain thesauri (Becker et al. 2009)
• Technique for automatically refactoring from action-noun style to labels in verb-object style
(Leopold, Smirnov, and Mendling 2012)
• Activities with similar labels provide potential refactoring opportunities
(Dijkman et al. 2011)
03.09.2012 rBPM 2012, Tallinn 24
Redundant Process Fragments Candidates for Reuse • PMS3: Redundant Process Fragments
o Redundant process fragments can be commonly found in existing process models
(Weber et al. 2011)
o More than 560 clones in the SAP reference model (Dumas et al. 20xx)
03.09.2012 rBPM 2012, Tallinn 25
Extracting Sub Processes • Method for automatically detecting exact clones,
which can be extracted to sub processes (Dumas et al. 20xx)
Similar process fragments as potential refactoring opportunities (Dijkman et al. 2011) o Exact matches o Similar fragments, whereby some activities
only appear in one fragment o Same activities, but different business objects
03.09.2012 rBPM 2012, Tallinn 26
Extracting Sub Processes • Method for automatically detecting exact clones,
which can be extracted to sub processes (Dumas et al. 20xx)
Similar process fragments as potential refactoring opportunities (Dijkman et al. 2011) o Exact matches o Similar fragments, whereby some activities
only appear in one fragment o Same activities, but different business objects
03.09.2012 rBPM 2012, Tallinn 27
Large Process Models are Difficult to Understand, Maintain, and to Reuse • PMS4: Large Process Models
o Process models with several hundred activities are not
uncommon (Soto et al. 2008)
o Large process models tend to comprise more formal flaws than smaller ones (Mendling et al. 2008)
03.09.2012 rBPM 2012, Tallinn 28
Automatically Extracting Sub processes • Method for automatic modularization of business
process models (Reijers et al. 2011)
• Method for the automatic labeling of process models (Leopold et al. 2011)
03.09.2012 rBPM 2012, Tallinn 29
Automatically Extracting Sub processes • Method for automatic modularization of business
process models (Reijers et al. 2011)
• Method for the automatic labeling of process models (Leopold et al. 2011)
TEST DRIVEN MODELING fostering understandability, maintainability, and reuse of declarative processes
03.09.2012 rBPM 2012, Tallinn 31
• Is the model really doing what it is expected to do?
Test Driven Modeling Comprehension Support Fostering Reuse
Comprehension
Modeling Reconciliation
03.09.2012 rBPM 2012, Tallinn
Interactions between Constraints Make Declarative Processes Difficult to Understand
Must be executed at least once
Must be first activity executed
Activity
Complete paper writing must be executed before Format to instructions
Must be executed once
Get acceptance must not be followed by Work on revision
Each execution of write response letter must be preceded by read reviews for revising paper
(Zugal et al. 2012b)
03.09.2012 rBPM 2012, Tallinn 33
Interactions between Constraints Make Declarative Processes Difficult to Understand
Must be executed at least once
Must be first activity executed
Activity
Complete paper writing must be executed before Format to instructions
Must be executed once
Get acceptance must not be followed by Work on revision
Each execution of write response letter must be preceded by read reviews for revising paper
03.09.2012 rBPM 2012, Tallinn 34
Interactions between Constraints Make Declarative Processes Difficult to Understand
Must be executed at least once
Must be first activity executed
Activity
Complete paper writing must be executed before Format to instructions
Must be executed once
Get acceptance must not be followed by Work on revision
Each execution of write response letter must be preceded by read reviews for revising paper
03.09.2012 rBPM 2012, Tallinn 35
Test Driven Modeling
Central Concept Test Cases: allows to
specify behavior the process model must exhibit or prohibit Test Case consist of
Execution Trace (1) Set of assertions (2), (3), (4)
(Zugal et al. 2012a)
03.09.2012 rBPM 2012, Tallinn 36
Test Driven Modeling Suite
(Zugal et al. 2012a)
03.09.2012 rBPM 2012, Tallinn 37
Test Driven Modeling
Makes information explicit that is only implicit in the process model, thus fostering understandability by supporting computational offloading
Allows for testing execution traces in an automated way
Test cases constitute an executable documentation
Reduces mental effort, speeds up changes, and reduces errors (Zugal et al. 2011, Zugal et al. 2012c)
LITERATE PROCESS MODELING
fostering understandability, maintainability and reuse by intervening text and model
03.09.2012 rBPM 2012, Tallinn 39
• What is the model doing and why have certain design decisions been taken?
Literate Process Modeling Comprehension Support Fostering Reuse
Comprehension
Modeling Reconciliation
03.09.2012 rBPM 2012, Tallinn 40
Literate Process Modeling
Supports domain experts and system analysts during model comprehension by flexibly interlinking textual descriptions and formal process model Fosters model understandability Provides a documentation of the process model Facilitates model comprehension as well as
model changes by documenting design decisions
(Pinggera et al. 2012b)
(Pinggera et al. 2012b) 41
Literate Process Modeling
SUMMARY
wrap up
03.09.2012 43
Conclusions
• Quality a necessary precondition for reuse • Creating, maintaining, and reusing process fragments
is a process in itself, which requires support
Comprehension
Modeling Reconciliation Process Model Refactoring
Test Driven Modeling Literate Process Modeling
03.09.2012 rBPM 2012, Tallinn 44
References • J. Becker, P. Delfmann, S. Herwig, L. Lis, A. Stein: Towards Increased Comparability of Conceptual Models –
Enforcing Naming Conventions through Domain Thesauri and Linguistic Grammars. In: Proc. of the 17th European Conference on Information Systems (2009).
• S.N. Cant, D.R. Jeffery and B. Henderson-Sellers: A conceptual model of cognitive complexity of elements of the programming process. Information and Software Technology 37 (1995) 7, pp. 351-362.
• R.M. Dijkman, B. Gfeller, J.M. Küster, H. Völzer: Identifying refactoring opportunities in process model repositories. Information & Software Technology 53(9): 937-948 (2011)
• M. Dumas, L. Garcia-Banuelos, M. La Rosa, R. Uba. Fast Detection of Exact Clones in Repositories of Business Process Models. Information Systems (to appear).
• D. Fahland: Number of ways students name an activity in a process model. http://dirksmetric.wordpress.com/2012/08/17/number-of-ways-student-name-an-activity-in-a-process-model/, [accessed on 2012/08/12]
• A. Hallerbach, T. Bauer and M. Reichert: Capturing Variability in Business Process Models: The Provop Approach. Journal of Software Maintenance and Evolution: Research and Practice 22 (2010) 6–7, pp. 519–546.
03.09.2012 rBPM 2012, Tallinn 45
References • Larkin, J.H., Simon, H.A.: Why a Diagram is (Sometimes) Worth Ten Thousand Words. Cognitive Science 11 (1987)
65-100
• H. Leopold, J. Mendling, H.A. Reijers: On the Automatic Labeling of Process Models. CAiSE 2011: 512-520
• H. Leopold, S. Smirnov, J. Mendling: On the refactoring of activity labels in business process models. Inf. Syst. 37(5):
443-459 (2012)
• J. Mendling: Metrics for Process Models: Empirical Foundations of Verification, Error Prediction and Guidelines for
Correctness, Springer, 2008.
• J. Mendling: Empirical Studies in Process Model Verification. Transactions on Petri Nets and Other Models of
Concurrency II, Springer, 2009, pp. 208–224.
• G. Miller: The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information.
Psychological Review 63 (1956), pp. 81-87.
• J. Mendling, H.A. Reijers and J. Recker, Activity Labeling in Process Modeling: Empirical Insights and
Recommendations, Information Systems 35 (2010) 4, pp. 467-482.
03.09.2012 rBPM 2012, Tallinn 46
References • J. Mendling, H.M.W. Verbeek, B.F. van Dongen, W.M.P. van der Aalst and G. Neumann: Detection and Prediction of
Errors in EPCs of the SAP Reference Model, Data & Knowledge Engineering 64 (2008) 1, pp. 312-329.
• F. Paas, J.E. Tuovinen, H. Tabbers, P.W.M.V. Gerven: Cognitive Load Measurement as a Means to Advance Cognitive
Load Theory. Educational Psychologist 38 (2003) 63-71.
• D.L. Parnas: On the Criteria to be Used in Decomposing Systems into Modules. Communications of the ACM 15
(1972) 1053-1058.
• J. Pinggera, T. Porcham, S. Zugal and B. Weber: LiProMo-Literate Process Modeling. In: Proc. CAiSE Forum ’12, pp.
163–170, 2012.
• J. Pinggera, S. Zugal and B. Weber: Investigating the Process of Process Modeling with Cheetah Experimental
Platform. In: Proc. ER-POIS ’10, pp. 13–18, 2010.
• J. Pinggera, S. Zugal, M. Weidlich, D. Fahland, B. Weber, J. Mendling and H. Reijers: Tracing the Process of Process
Modeling with Modeling Phase Diagrams. In: Proc. ER-BPM ’11, pp. 370–382, 2012.
03.09.2012 rBPM 2012, Tallinn 47
References • J. Pinggera, P. Soffer, S. Zugal, B. Weber, M. Weidlich, D. Fahland, H. Reijers and J. Mendling: Modeling Styles in
Business Process Modeling. In: Proc. BPMDS ’12, pp. 151–166, 2012.
• H.A. Reijers, J. Mendling, R.M. Dijkman: Human and automatic modularizations of process models to enhance their
comprehension. Inf. Syst. 36(5): 881-897 (2011)
• M. Soto, A. Ocampo and J. Munch: The Secret Life of a Process Description: A Look into the Evolution of a Large
Process Model, In: Proc. ICSP'08, 2008, pp. 257-268.
• Sweller, J.: Cognitive load during problem solving: Effects on learning. Cognitive Science 12 (1988) 257-285.
• Sweller, J., Chandler, P.: Why Some Material Is Difficult to Learn. Cognition and Instruction 12 (1994) 185-233.
• B. Weber and M. Reichert: Refactoring Process Models in Large Process Repositories. In: Proc. CAiSE'08 (2008), pp.
124-139.
• B. Weber, M. Reichert, J. Mendling and H.A. Reijers: Refactoring Large Process Model Repositories.. Computers and
Industry 62(2011) 5, pp. 467-486.
03.09.2012 rBPM 2012, Tallinn 48
References • S. Zugal, J. Pinggera and B. Weber: The Impact of Testcases on the Maintainability of Declarative Process Models.
In: Proc. BPMDS ’11, pp. 163–177, 2011.
• S. Zugal, C. Haisjackl, J. Pinggera and B. Weber: Empirical Evaluation of Test Driven Modeling. International Journal
of Information System Modeling and Design (accepted), 2012
• S. Zugal, J. Pinggera and B. Weber: Toward Enhanced Life-Cycle Support for Declarative Processes. Journal of
Software: Evolution and Process 24(3):285–302, 2012
THANK YOU FOR YOUR ATTENTION
http://bpm.q-e.at @bpm_qe