This document summarizes a study on software quality concerns in the Italian banking sector. The study found that bank information systems have complex architectures that are difficult to evolve and maintain over time. Software quality is decreasing as demands for new features increase while budgets remain limited. A key finding was that software architecture quality emerged as an important but missing dimension when mapping factors to existing quality standards. The researchers propose a new "meta quality dimension" model that considers how quality standards interact rather than viewing them in isolation.
1. Software Quality Concerns
in the Italian Bank Sector:
the Emergence of a Meta-Quality Dimension
Daniel Russo*, Paolo Ciancarini, Tommaso Falasconi, Massimo Tomasi
Buenos Aires, Argentina. May, 2017
ICSE– SEIP Track
2. Why?
• Issues from banking managers’ community:
ü Perceived decreasing of software quality in the
banking sector
ü Strongly stratified architectures
ü Difficult evolution & maintenance of banking
applications
ü Difficult reverse engineering
ü …
• Minimal literature
3. What?
• Mixed Methods research:
– Pragmatism
ü Inductive: software quality concerns
identification
ü Deductive: concerns validation
• Delphi-like study:
ü Traditional Delphi
ü + Survey
4. How?
• Vertical Study:
ü Sector specific
ü Country specific
• Expert pool selected by industrial co-
authors:
ü Deep insight within the Italian banking
industry
7. Findings at a glance
• Bank Information Systems are characterized
by complex stratified architectures
• Sinking software quality
• Increasing requests for new functionalities
• Limited IT budgets
ü Do more with less?
• Poor IT business alignment
ü C*O rarely understand CIOs
8. Factors and results I
# Question “Agree” and
“Strong Agree”
Avg Score
1 Module interfaces complexity 96% 1,95
2 Interfaces architecture complexity 96% 2,47
3 Custom software quality 77% 1,15
4 Increase of maintenance costs 82% 1,58
5 Quality vs. Time & Budget 79% 1,57
6 Quality vs. System analysis 79% 1,00
7 System analysis vs. Documentation 87% 1,70
8 Documentation vs. Time & Budget 84% 1,46
9. Factors and results II
# Question “Agree” and
“Strong Agree”
Avg Score
9 New packages functionalities vs.
Complexity
83% 1,58
10 Packages vs. Documentation 82% 1,30
11 Packages documentation vs.
System analysis
77% 1,09
12 Application & Maintenance contracts
vs. Documentation.
77% 1,15
13 Non-Italian applications vs. Quality
& Maintainability
76% 1,29
14 Italian applications vs. Quality &
Maintainability
75% 0,92
15 Measurement of software quality 70% 0,93
10. Discussion
• Factor mapping within the ISO/IEC
25010:2011 standard:
ü Product Quality
ü Quality in Use
• But… something was missing
Software Architecture
11. Discussion
• Direct or indirect reference to
architecture
• Stressing architectural concerns
• Relation between quality and
architecture dimension
• Meta Quality Dimension:
ü Taking standards not as silos
ü Quality is given by the interaction
13. Conclusions
• Short term view
• Relations among different software
dimensions
• Standards should not be taken as silos:
ütheir interaction is important
14. Future works
What about interpreting information
system’s quality with:
Quality
ProcessArchitecture
• Mixed
Methods
• Process
Mining
15. Daniel Russo
Department of Computer Science & Engineering – DISI
Italian National Research Council – CNR-ISTC
Consorzio Interuniversitario Nazionale per l’Informatica – CINI
daniel.russo@unibo.it
www.unibo.it