Thoughts on Teaching Software Quality

Engineering

Witold Suryn
Department of Electrical Engineering
École de Technologie Supérieure Montréal, Canada
[email protected]

ABSTRACT
The article presents an overview of the subject of Software Quality
Engineering (SQE) education. Four different perspectives are taken into
account: why to teach SQE, how the subject is being taught today, what
support teachers have to teach SQE and how could the Software Quality
engineer be educated. The latest trends in methods and tools pertinent to the
domain are also presented.

1. Introduction
Quality becomes a critical attribute of a software product since its absence
results in dissatisfied users, loss of money, and – in the most critical
situation – in lost lives. Increasing recognition of the importance of software
quality makes the software engineering “center of gravity” shift from
creating a technology solution toward a quality solution. Development
organizations confronted with such an approach are, in general, not
prepared to deal with it since their engineers too often are not adequately
educated. And if such knowledge exists, it is still more of an empirical
nature than resulting from a formal educational process. This reality makes
the role played by Software Quality Engineering (SQE) education to change
– or sometimes “about to change” - from “Cinderella” to a full member of
Software Engineering Education family. This paper presents a quick look at
today’s situation in Software Quality Engineering education from four
different points of view: why should SQE be taught, how is SQE being
taught, what has an SQE teacher for his disposition and what and how could
the future SQE engineer be educated. This paper has no ambition to cover
the subject to its fullest extent, seeking rather a wide reaction from the
community of academic teachers and researchers.
2. Why should Software Quality Engineering be
taught

Applying a simplistic mathematic formula the answer could be presented in

the following manner:

RISK= F (1/quality)
or,

The increased quality mitigates the risk

The ways in which the risk resulting from missing quality can manifest
itself are numerous, however for a user the mostly recognized or
experienced categories could be:
• Discomfort (ex: famous blue screens)
• Social or professional losses (information/data, time, money,
job etc.)
• Loss of health or life (ex: over-X-rayed patients in one of
Bristol hospitals)

It is rather obvious that our (i.e. users’) dream about a bug-free, highest
level quality solution for a small price will remain a dream for next n years,
but how big this n will be, depends on the level of SQE education and
knowledge that will find its way to software development companies. Most
probably there is a chance to buy one day a software Rolls Royce but let’s
not forget that for car manufacturers it took a hundred years to obtain a
sustainable high quality. Do we have to wait that long?
Missing quality takes its toll also from a developer/supplier even if his
position is still so privileged that better word would be “untouchable”. Until
recently the biggest risk resulting from low or missing quality of software
that a supplier could face was loosing a customer, money or market. An
individual user was entirely helpless against a supplier whose low quality
software created problems. But this situation evolves in the direction where
a user or a group of users can seek compensation of their losses through
legal actions [1]. Such an option adds two other risks to the list of supplier’s
risks:
• A court trial with a full range of consequences, and, when worse
comes to the worst,
• An imprisonment.
It is then understandable why SQE receives a continuously increasing
attention from software development companies. One of very positive
consequences of the recent software-related legal evolution is an observable
intensification of industry-funded joint research projects aimed to develop
science and best practices in SQE [2, 3].
To conclude one could propose the following statement: Software Quality
Engineering should be taught because the time when both customers and
suppliers will be fighting for best quality is knocking to our door.

3. How is Software Quality Engineering being

taught
How in reality SQE is being taught today only a teacher can answer so as
many theories and practices as many teachers. Such a “liberty” is usually
symptomatic to domains of knowledge that would obtain maximally CMM
level “one-and-a-half” on the maturity scale. And that is true; SQE is
immature, but not entirely. There are some considerable efforts aiming to
create a structured approach to the SQE body of knowledge and teaching
practices. First serious, professional efforts can be dated to early ‘90s [4]
however in this paper only two most recent projects will be discussed.
Project 1: “Guidelines for Software Engineering Education” Version 1.0.
Software Engineering Institute (USA), 1999. [5]
The document proposes the following structure of the body of knowledge of
Software Engineering education:
z Core Area:
¾ Software Requirements
¾ Software Design
¾ Software Construction
¾ Software Project Management
¾ Software Evolution
z Foundations Area
¾ Computing Fundamentals
¾ Human Factors
¾ Application Domains
z Recurring Area (Components in the Recurring Area are threads
that occur through all of the Core Area components):
¾ Ethics and Professionalism
¾ Software Processes
¾ Software Quality
¾ Software Modeling
 ¾ Software Metrics
¾ Tools and Environments
¾ Documentation
z Supporting Area that includes, but is not limited to general
education, mathematics, […] and engineering

The proposed structure is well organized and offers a good support for those
responsible for devising education curricula, however from SQE
perspective of year 2003 this solution is hardly acceptable. Putting Software
Quality in Recurring Area (refer to the explanation above) allows “gluing”
this component freely wherever one could please. Moreover, it is practically
infeasible to “glue” such a large component in one piece to another
component, so in order to allow “gluing” the component has to be
partitioned. How and following what logic?

Project 2: Software Engineering Education Knowledge (SEEK), IEEE

Computer Society, ACM.
After SEEK, First Draft [6]:
“The education knowledge area group is responsible for defining and
documenting a software engineering education body of knowledge
appropriate for guiding the development of undergraduate software
engineering curricula. This body of knowledge is called Software
Engineering Education Knowledge or SEEK. The pedagogy focus group is
responsible for using SEEK to formulate guidance for pedagogy as well as
course and curriculum design to support undergraduate software
engineering degree programs. The initial selection of the SEEK areas was
based on the Software Engineering Body of Knowledge (SWEBOK)
knowledge areas”

This innovative project with ambitions to become an international reference

source for Software Engineering education programs as the first recognizes
Software Quality as an individual knowledge area (KA) that requires
individual teaching thread.

Phase: SEEK, First Draft, August 2002 [6].

Proposing Software Quality as the separate KA makes SEEK the most
modern path leading to SQE maturity, but in spite of that the project is not
free from critic. Quoting from [8]:
z 17 hours of core contact (hours of teaching) dedicated to all the
content of Software Quality component in comparison with 250
hours for Software Engineering Fundamentals seems to be
seriously underscored
 z The presence of Software Quality Standards indicates that
standards have finally become part of modern software engineering
education; however the documents proposed as seminal do not
always represent the state-of-the-art.
z There are subjects that did not yet find their place in SEEK’s
Software Quality KA, as f.e. Software Quality Management and
Evaluation or Software Quality Implementation.
For more details on critic of SEEK, First Daft refer to [8].

Phase: SEEK Second Draft, December 2002 [7]

The critical opinions about First Draft expressed by researchers and teachers
participating in the international conference on Software Technology and
Engineering Practices (STEP) in Montreal, Canada in 2002 have been taken
into account during redaction of Second Draft of SEEK project. The
outcome however still leaves something to desire as the improvements seem
to be rather cosmetic (ex: the number of teaching hours was increased from
17 to 21). For more details refer to [7].
In summary – SQE education is still being led by “heroes” (CMM level
one) but what makes it reach the level “one-and-a-half” is SEEK. The only
thing that has to be done now is to make SEEK also mature.

4. What can a teacher use to teach Software

Quality Engineering
In year 2003 even scientific “heroes” do not stay without at least some
support. In case of SQE this support is present on all crucial levels: theory,
practice, tools and experience, but unfortunately is not very reach.

On the level of theory two major projects offer some support to an SQE
teacher: SEEK and SWEBOK.
SEEK, as it was presented in the previous chapter, is an ambitious project
with rather good provisions for success but still relatively far from
completeness and stability.

SWEBOK (Software Engineering Body of Knowledge) [9] however is the

project with over 5-year history, realized with participation of over 500
specialists worldwide aiming to create internationally recognized guide to
Software Engineering Body of Knowledge. The results of this project were
found so satisfactory that several educational institutions took SWEBOK as
the basis for modernizing undergraduate Software Engineering programs
[10, 11].
For SQE education SWEBOK offers:
z Recognition of Software Quality as a separate Knowledge Area
z Recommended Software Quality KA contents (several KAs are
being under thorough revision what already resulted in proposed
enhancements and improvements [12, 13]).
z Recommended breakdown of topics (Figure 1)
z Seminal references, including standards
z Further reading

Software Quality

Purpose and Activities and Measurement

Software Quality
Planning of SQA Techniques for Applied to SQA
Concepts
and V&V SQA and V&V and V&V

Measuring the Common Planning Static Techniques Fundamentals of

Value of Quality Activities Measurement
Dynamic Techiques
ISO 9126 Quality The SQA Plan Measures
Description
Other SQA and
The V&V Plan V&V Testing Measurement
Dependability Analysis
Techniques
Special Types of
Systems and Quality Defect
Needs Characterization

Additional Uses
of SQA and
V&V data
Figure 1. SWEBOK’s Software Quality Knowledge Area breakdown of
topics

On the level of practice an SQE teacher is supported by several elements

that represent so sought maturity and stability. Among these important
elements are:
z Quality models: McCall, Boehm, Dromey [14], ISO/IEC 9126 [15]
z Software Product Quality Measurement and Evaluation standards
o ISO/IEC 9126 [16, 17,18]
o ISO/IEC 25000 SQuaRE (next generation of ISO 9126)
[19]
z Software Product Quality support processes standards
o ISO/IEC 14598 [20]
o ISO/IEC 25000 SQuaRE (next generation of ISO/IEC
14598) [19]
z Software Measurement standards
o ISO/IEC 15939 [21]
o ISO 19761 COSMIC-FFP [22]
z Software Life Cycle Processes standards
o ISO 12207 : 1995 (plus amendments 2002) [23]
o ISO 15288 : 2002 [24]

On the level of tools and methods the possible support is reduced to a

few proprietary products remaining either in stage of a prototype or
being under reconstruction. Despite of their actual status these tools
demonstrate a vivid interest of academia and industry centers in
developing non-manual means of evaluating the quality of software.

These tools are briefly presented below:

z Software Quality Assessment Exercise (SQAE) – Robert A.

Martin and Lawrence H. Shafer from MITRE Corp USA. April
1996 [25]
¾ Based on quality model built on the combination of
Boehm, McCall and Dromey
¾ Evaluates only static quality attributes (4 areas, 7 factors,
>100 questions)
¾ Toolkit that is being considered as requiring some
modifications [2]
¾ Remaining the subject of the research of graduate students
at ÉTS, verifying the possibility to migrate SQAE to
ISO/IEC 9126 [3]
z Multi-layered Customizable Software Quality Model – Elli
Georgiadou from Middlesex University, UK, 2003 [26]
¾ Based on ISO/IEC 9126
¾ Innovative techniques as Composite Features Diagram
and Kiviat Diagram
¾ Profiler tool in development, not available publicly
 z QEST Evaluation Toolkit – Alain Abran from École de
Technologie Supérieure, Montréal, Canada and Martin Kunz, Otto-
von-Guericke University Magdeburg, Germany. 2003 [27]
¾ Based entirely on ISO/IEC 9126 with very good graphical
interpretation
¾ Non-profit, not for sale
¾ For researchers primarily
¾ Accessible anytime-anywhere (Web-Internet)
¾ Core part only

The theme of the experience support to an SQE teacher leaves a bit of a

bitter arrière-goût (in French – after taste) on the lips. The community of
SQE researchers and teachers makes a huge intellectual power with great
experience gathered over the years, but:
z usually unpublished or published only fractionally
z kept as notes or course materials
z too often not shared or exchanged in a larger scale
z unevenly distributed (droplet-type knowledge)
z not synchronized

In effect of the above walkthrough of the teaching support to SQE a few

conclusions to this chapter could be proposed:
z Teaching SQE is still the role for a “hero”, however not without
some support.
z Software Quality Engineering may require its own Body of
Knowledge named for example SQEBOK (the author reserves for
himself the title of the ownership of this name) to advance the
maturation process,
z As there is no SQEBOK, can we, SQE researchers and teachers
become a kernel of this project?

5. What and how the Software Quality Engineer

could be educated
The response to this question is hidden in the word “engineering”. A
Software Quality engineer, as any other engineer is supposed to leave in
certain moment his alma mater and go to the industry to help it improve and
evolve. In the light of such a hypothesis an SQ engineer’s education should
allow him to implement, measure, evaluate and improve the quality of
software throughout its entire lifecycle. Such a vast knowledge requires a
structured approach that allows for systematical building the acquired
experience that remains, if possible, in accord with industrial practices. A
possible structure could consist of, but not be limited to three major (core)
components (this proposition focuses only on core areas of knowledge, for
example: in Quality Requirements component the techniques of
requirements elicitation are taken as prerequisites):
z Quality requirements
z Quality measurement and evaluation instruments
z In-lifecycle quality implementation

Knowledge acquired in Quality requirements component should primarily

allow an SQ engineer to extract quality requirements from high-level
stakeholder’s requirements and then to decompose them into lower level
categories of quality requirements down to corresponding quality measures
(Figure 2) [28].

Stakeholders’
Requirements

Functional & non- Quality Requirements

Functional
Requirements

Operational
Quality in Use Quality

Impact? External Quality

Internal Quality

Figure 2. Quality requirements decomposition model (Suryn et al [19])

The proposed decomposition model is based on quality model from

ISO/IEC 9126 in conjunction with TL 9000 standard [15, 29, 30] where
“the requirements of Quality in Use contribute to specifying External
Quality requirements, which in turn contribute to specifying Internal
Quality requirements” [28]. The model is well documented so relatively
easy to learn and use, but static. For a proficient SQ engineer it is important
to know not only what but also how. This approach is addressed by the
model for practical process of defining and controlling quality requirements
from Fig.3.

Stakeholders’
Requirements

Operational
Quality Quality
Requirements Requirements

Quality
in Use
Requirements

External Quality
Requirements

Internal Quality
Requirements

Legend:
Solid arrows ask the question HOW
Boxes ask the question WHAT (choices)
Dashed arrows ask about TRACING

Figure 3. Practical process of defining and controlling quality requirements

(Suryn et al. [19])

In this model solid arrows indicate the path of questions “how” that have to
be asked when executing requirements decomposition, boxes ask questions
“(into) what” to help define requirements that result from decomposition
process and dashed arrows indicate the path of the subject of traceability of
requirements.
The theoretical and practical knowledge of responses to the above questions
should allow the SQ engineer to professionally manage the process of
quality requirements analysis and definition.
The Quality measurement and evaluation instruments component has as the
objective to educate the SQ engineer in existing quality models and
measures that allow measuring chosen quality attributes of software, and in
processes that support this activity. Again, these two elements are well
documented and can be easily taught, but also are static. The engineering
part of the knowledge, which is mapping of these instruments into software
lifecycle phases (Fig.4), is much less documented and requires further
research. Such a situation may however turn more promising from
educational process point of view, as usually the knowledge acquired
actively stays longer than this acquired through lectures.

Stakeholder
Requirements
Definition

Requirements
Analysis Quality
Instruments

Architectural
Design

Implementation Mapping

Integration

Verification
Evaluation
Process
Transition

Validation

Operation &
Maintenance

Figure 4. Mapping software quality measurement and evaluation to phases

of software lifecycle (Suryn et al. [19])
The last component, In-lifecycle quality implementation results from two
previous components where the basic knowledge has been acquired. The
implementation now requires the practical application of this knowledge in
course of developing the software (Fig.5).

process software product effect of software

product

influences influences influences

internal external quality in
process quality
quality use
quality attributes
depends on attributes depends on depends on attributes contexts
of use

process internal external quality in use

measures measures measures measures

Product Development
and Maintenance

SW Product Quality
Requirements

Product
in
Use

Figure 5. Quality implementation challenge model (Suryn et al. [19])

The education process in this component should allow the SQ engineer to
accept the responsibility in the industrial environment not only to identify
the correspondence between software lifecycle and quality lifecycle or to
define and control quality requirements but first of all to know how to
implement, measure, evaluate and improve the quality in all phases of
software lifecycle. This role requires and interactive co-operation with
development teams and engineers.
But this part is still emerging and requires more data coming from industry
and more research affected by the academia.

6. Conclusion
Software Quality Engineering enters now its very active phase of the
research, creating best practices and industry application. The academic
teachers posses enormous knowledge, experience and will to discover new
areas and pass it all to students who will convey the new science and best
practices to the industry. There are new projects in the domain of Software
Engineering body of knowledge and education together with works of
International Organization for Standards (ISO) that offer a considerable
support to teachers. All that indicates a very promising future for Software
Quality Engineering, but to achieve the maturity of this domain both the
academia and industry have to invest more in research in many incoming
years.

References
1. Abreu E., M., “Consumer groups target software”. Reuters, June 14,
2002
2. Halde M., Martin R., Beauchemin P., “Évaluation de la qualité d'un
logiciel à l'aide de la méthode SQAE de MITRE”. École de technologie
supérieure, 2003
3. Côté M-A., Ktata O., Azzouz R., “Migration et Adaptation du Software
Quality Assessment Exercise de Mitre Corp”. École de technologie
supérieure, 2003
4. Shaw M. “We Can Teach Software Better”. Computing Research News,
4, 4 September 1992 (pp. 2, 3, 4, 12)
5. Bagert D., J., et al. “Guidelines for Software Engineering Education”
Version 1.0. Technical report cmu/sei-99-tr-032 esc-tr-99-002. SEI October
1999
6. Sobel A., E., K., “Computing Curricula - Software Engineering
Volume”. Software Engineering Education Knowledge (SEEK), First Draft.
IEEE Computer Society and ACM. August 28, 2002
7. Sobel A., E., K., “Computing Curricula - Software Engineering
Volume”. Software Engineering Education Knowledge (SEEK), Second
Draft. IEEE Computer Society and ACM. December 6, 2002
8. Vliet H. van, Suryn W. et al. “Thoughts on Software Engineering
Knowledge, and how to organize it”. IEEE Computer Society, 2003.
Prepared for May 2003.
9. Abran, A., Moore, J.W., Bourque, P., Dupuis, R., Tripp, L. “Guide to the
Software Engineering Body of Knowledge – Trial Version”. IEEE
Computer Society, Los Alamos, 2001
10. Frailey D. et al. “Using SWEBOK for Education Programs in Industry
and Academia”. Southern Methodist University and Raytheon Company,
2002
11. Suryn W. et al. “Amélioration de la structure du programme
baccalauréat en Genie Logiciel”. ”. École de technologie supérieure,
September 2002
12. Suryn W.,, Robert F., Abran A., Bourque P., Champagne R., “
Experimental Support Analysis of the Software Construction Knowledge
Area in the SWEBOK Guide (Trial Version 1.0)”. IEEE Computer Society,
2003. Prepared for May 2003.
13. Palza E. “Analyse du contenu du chapitre « Requirements » dans
SWEBOK Version 0,95”. École de technologie supérieure, Decemeber
2001.
14. Vliet H. van., Software Engineering, Principles and Practice , Second
Edition. Wiley & Son,s 2001.
15. ISO/IEC 9126 – Software Engineering – Product Quality - Part 1:
Quality Model. 2001
16. ISO/IEC 9126 – Software Engineering – Product Quality - Part 2:
External Metrics. Planned for 2003.
17. ISO/IEC 9126 – Software Engineering – Product Quality - Part 3:
Internal Metrics. Planned for 2003.
18. ISO/IEC 9126 – Software Engineering – Product Quality - Part 4:
Quality in Use Metrics. Planned for 2003.
19. Suryn W., Abran A., “ISO/IEC SQuaRE. The second generation of
standards for software product quality”. Submitted to QSIC 2003 – Third
International Conference on Quality Software Beijing, China, September
25-26, 2003
20. ISO/IEC 14598 Information Technology - Software Product
Evaluation, Parts 1-6. 1999-2001
21. ISO/IEC 15939 - Software Engineering - Software Measurement
Process. 2002
22. ISO/IEC 19761 Software Engineering – COSMIC FFP: A functional
sizing method, 2003
23. ISO 12207 - Information technology - Software Engineering - Software
Life-Cycle Processes, 1995
24. ISO/IEC 15288 - Information Technology - Life Cycle Management -
System Life Cycle Processes. 2002
25. Martin R., A., Morrison S., A., Shafer L., H., “Providing a Framework
for Effective Software Quality Assessment: First Step in Automating
Assessments”. The MITRE Corporation, April 1996
26. Georgiadou E. et al. “Towards a Multi-layered Customizable Software
Quality Model”. Proceedings of SQM2003
27. Kunz M. “The prototypical web-based implementation of the QEST
model”. Otto-von-Guericke University Magdeburg, Germany. 2003
28. Suryn W. et al. “Software Product Quality Practices Quality
Measurement and Evaluation using TL9000 and ISO/IEC 9126”. IEEE
Computer Society, 2003. Prepared for May 2003.
29. TL9000 Quality Management System Requirements Handbook,
Release 3.0, QuEST Forum 2001
30. TL9000 Quality Management System Measurements Handbook,
Release 3.0, QuEST Forum 2001