Upload
henry-muccini
View
484
Download
4
Embed Size (px)
Citation preview
Università degli Studi dell’Aquila
1
A Study on Group Decision-Making in Software Architecture
Smrithi Rekha V.Amrita Vishwa, Vidyapeetham, [email protected]
Henry Muccini, Ph.D. University of L’Aquila, Italy [email protected]
@muccinihenry, henrymuccini.com
Presented @ WICSA 2014, Sydney, Australia
Preamble2
Multiple stakeholders are involved
Each with different concerns and goals
Decision Making
Tech. Stakeholders
Customers
…
…Business
Final User
Architects
42010:2011
SA decision-making is a group process
Implications3
SA results from a Group Decision Making (GDM) process
While methods and tools have been proposed to record ADD and their rationale
More can be done for handling GDM
in Software Architecture
Group Decision Making4
Three decades of research on group decision making in the business domain
GDM Research Perspectives
Processes and Methods Impact of factors like size, diversity, roles, tasks
Challenges
Comparative Studies: Various methods, Individual vs Group Issues: Groupthink, Group
Shift
Conflict Resolution
Process Enhancement
Pros and Cons
GDM has been studied from multiple perspectives that includes Psychology,
Organizational Behavior, Operations Research and Economics
Picture taken from http://apprentiperpetuel.blogspot.com.au/
Aim of this study5
how ◄practitioners► make group decisions in architecting software systems
how ◄state of the practice► GDM in SA relates to ◄state-of-the-art► GDM techniques
◄challenges►companies face when making architecture-related group decisions
To understand
RQ1
RQ2
RQ3
Study Plan6
Online questionnaire
38 questions
Q1-Q14: personal and company info
Q15-Q25: undertaken SA group decision-making process
Q26-Q31: SOTP vs SOTA SA GDM process in industry
and methods available in literature
Q32-Q36: challenges
30 respondents
23 practitioners involved in SA GDM
7 academics knowning about SA GDM in industry
64% architects21% sw development15% top management
3-16 years of experience
5-10 people involved in decision making 21 different macro-
roles represented 80%-20% distributed-
colocated Distributed:
54% uses only synch comm. media
4% asynch only
50% uses tools to support the ADD process
Profile of participantsQ15-Q25:undertaken SA group decision-making process
7MAIN FINDINGS (1/3)
Drivers: time, cost and quality Decision pattern: group
discussion -> if no agreement, a leader/small group decides (35%)
57% uses a mix of GDM methods
Stakeholders are given different priority: seniority (55%), business, political and technological factors (25%), individuals (10%)
Q15-Q25 (cont.)Q26-Q31: Comparing SOTP with SOTA
8MAIN FINDINGS (2/3)
how practitioners arrive at a consensus
Conflicts: may occur due to socio-cultural reasons technical reasons cost
Q32-Q36: Challenges faced in GDMQ32-36 (cont.)
9MAIN FINDINGS (3/3)
GroupThink appears evident!!
arriving atconsensus becomes
more important than individual opinions
Time-bounded decisions Improved documentations Tool support improvement
Peculiarities of GDM in SA Exploring into causes and
possible mitigation strategies for GDM issues
metrics for group decision quality
…
C. Miesbauer and R. Weinreich @ECSA 2013
D. Tofan, M. Galster, and P. Avgeriou @ ECSA 2013
Shahin, Liang, Khayyambashi, @WICSA/ECSA 2009
M. Nowak and C. Pautasso @ WICSA 2012
T. Al-Naeem, I. Gorton, M. Babar, F. Rabhi, and B. Benatallah @ICSE 2005
Suggestions Closely Related Work
10
Future Work
Contacts
If you are interested to this research, please
11
Stop by after the presentation
Contact me at [email protected]
Tweet @muccinihenry
Skype me at henry.muccini
Call me