Managing Software Process Evolution: Traditional, Agile and Beyond – How to Handle Process Change

345 Pages·2016·7.89 MB·English

by  Marco Kuhrmann

Marco Kuhrmann · Jürgen Münch
Ita Richardson · Andreas Rausch
He Zhang E ditors
Managing
Software
Process
Evolution
Traditional, Agile and Beyond –
How to Handle Process Change
Managing Software Process Evolution
ü ü
Marco Kuhrmann J rgen M nch
(cid:129)
Ita Richardson Andreas Rausch
(cid:129)
He Zhang
Editors
Managing
Software
Process
Evolution
–
Traditional, Agile and Beyond
How to Handle Process Change
123
Editors
MarcoKuhrmann Andreas Rausch
University of SouthernDenmark Clausthal University of Technology,
Odense Departmentof Informatics
Denmark Clausthal-Zellerfeld
Germany
JürgenMünch
Reutlingen University He Zhang
Reutlingen NanjingUniversity
Germany Nanjing
China
Ita Richardson
University of Limerick
Limerick
Ireland
ISBN978-3-319-31543-0 ISBN978-3-319-31545-4 (eBook)
DOI 10.1007/978-3-319-31545-4
LibraryofCongressControlNumber:2016934950
©SpringerInternationalPublishingSwitzerland2016
Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpart
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
orinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilar
methodologynowknownorhereafterdeveloped.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfrom
therelevantprotectivelawsandregulationsandthereforefreeforgeneraluse.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authorsortheeditorsgiveawarranty,expressorimplied,withrespecttothematerialcontainedhereinor
foranyerrorsoromissionsthatmayhavebeenmade.
Printedonacid-freepaper
ThisSpringerimprintispublishedbySpringerNature
TheregisteredcompanyisSpringerInternationalPublishingAGSwitzerland
Foreword
Whether we like it or not, the business software is a fashion industry where new
fadscomeandgo.Allofuscannameseveralofthemfromthepastyearsandshare
the war stories about attempting to fix problems with the newest fad at everyone’s
lips. Today some of these fads are called continuous deployment, lean startups,
DevOps,flowandanyoftheJapanesesoundingconceptsfromtheleanvocabulary
such as Kanban for some years now. So, who cares about software processes
anymore?
Back in the days of my software development in the 1990s in a 10-person
software boutique producing production planning software for Nokia factories, I
becameworriedaboutthequalityofoursoftwareandsharedmythoughtswiththe
chief architect. He made me a quality manager for the company and a few weeks
laterNokiaperformedanISO9001assessmentforourboutiqueandthesoftware.It
turned out that we deviated in all other points but at least we had an appointed
quality manager to worry about the results. Among others, we did not have pro-
cesses in place.
Later I made my Ph.D. in Software Process Improvement, became a certified
BOOTSTRAPandSPICEassessor,anddiscoveredthatifthereisawaytoimprove
performance, all sorts of process improvement endeavors are not the way forward.
As much as we’d like, the processes are not tangible artifacts, they do not deliver
software and when improved, we cannot quantify the results in order to determine
whether we are developing better software than yesterday.
The past decade and a half has still been mostly about the processes, methods,
and tools; but in a refreshing way. Agile kicked out process developers and
empowered developers to become concerned about the business of the company,
the actual software development. Agile grew rapidly outside of its origins from
small teams to larger teams, to organizations and to global software development.
Yes, we still agree to disagree what exactly agile is and what it means in different
contexts in precise terms. We have learned that each and every organization must
defineitforthemselves.Thishasproventobepainfulbutnecessarywhentheideas
of continuous improvement and learning organizations are nurtured.
v
vi Foreword
I am glad to say that the industry is moving forward, academics are reforming
the curriculums, practitioners are more involved than ever before, and improve-
ments or even improvement leaps are being made.
Thisiswherethebookatyourhandbecomesveryhandy.Itshowshowsoftware
processes evolve and are impacted in various contexts and how this can be man-
aged.Inaninterestingwaythebookcontributestotheerabeyondagile,whichwill
bewelcomedbymany.Also,Ishouldnotethatthebookdoesnotattempttoanchor
itself in one particular domain, context or business situation but offers different
perspectives, which will benefit readers with different backgrounds.
The editors have collected 15 chapters from authors that represent both the
academic and the practitioner standpoints. Some of the chapters are refreshingly
controversial and many provide also concrete guidance on how to make a lasting
impact.Allofthechapterscommunicatethesameunderlyingmessage,whichIfeel
is very important: “We must keep changing the way the software is being devel-
oped.” Iparticularly appreciated thesense ofurgency topushtheindustry forward
fromamultitudeofdifferentviewpoints.Whileallchaptershavetheirmerits,Iwill
highlight a few that caught my attention:
Anthony I. Wasserman (Chap. 1) presents the ultimate key question that has
puzzledusforanumberofyears:“Howmuchprocessisneededbyanorganization
for a particular project?”. In many ways he sets the stage for the remainder of the
book by arguing that there are only a few places left where high-ceremony (a.k.a.
traditional) processes are needed.
Andreas Rösel (Chap. 5) challenges the reader by presenting an approach to
guaranteeafailure whengiganticimprovementsshouldbesoughtbutwhenwe do
not dare to take the risk. Many of the anti-tactics he identifies should be an
eye-opener for software managers thinking about their improvement efforts.
Christian Prause and his colleagues’ (Chap. 8) chapter, on the other hand, is an
eye-opener for all those people who think software development is an easy
endeavor. They describe their environment insightfully: “Scientific missions have
no insurance; a second unit is never built. If the mission goal is not reached, for
whatever reason, there is no second chance.” Software that needs to operate for
decades in the outer space requires all the ceremonies invented but a smartway of
executing them.
Kai Petersen’s (Chap. 12) contribution should be an absolute read to all
researchersandpractitionersinvolvedinacademia–industrycollaboration.Inavery
practical way, he summarizes the literature on the topic but also adds his personal
adviceonhowtheresearchdonebytheacademicscandeliverthemaximumvalue
for the company.
Yli-Huumo and his colleagues (Chap. 15) demonstrate how changes in the
organization and the way the software is being developed do have a fairly direct
impact on the technical debt that hinders the company to progress and move for-
wardintheirdevelopment.Thereaderswillparticularlyenjoytheillustrativequotes
from the interviews in three large companies.
Foreword vii
ComingbacktowhereIstarted,itturnsoutthatsoftwareprocesseshavebecome
evermoretangibleandconcreteactionscanbetakenthatwillshowadifferencein
the bottom line. The editors have put together a book that does a splendid job in
fulfilling an evident gap in the current literature by shaping the state of the art in
software process evolution scene.
Whether you read this book from start to finish, or piecemeal your approach
iteratively, I am sure you will find this book as valuable as I did.
Trondheim, Norway Pekka Abrahamsson
January 2016 Professor of Software Engineering
Norwegian University of Science and Technology
PS:IforgottoconcludethestoryabouttheISO9001assessmentinthesoftwareboutiquewhereI
worked. After surviving the shock of being evaluated by Nokia, my company jumped in the
processwagon,eventuallyIdepartedtopursueacademicstudies,andthecompanyendedupbeing
recognizedbyInternationalQualityCrownawardin2008andArchofEuropeQualityawardin
2010.Thus,presentingahappyendingforoneprocessassessmentcase.
Preface
Imaginethishappenstoyou:Yourmanagertellsyou“Agileisthefuture!Let’sgo
Scrum.” He forces you to replace the existing development process with Scrum.
What would you do? Would you send your developers to a Scrum training course
immediately?
It is true that more companies are embracing agile as part of their development
process in order to increase speed, accelerate learning, and deliver value rapidly.
AndmanyofthesecompaniesareapplyingScrum.Butitisalsotruethatevolution
doesnotfollowtheprinciple:“Progressivedinosaursarethefuture!Let’sgobird.”1
Evolving the ways through which software-intensive products and services are
developedisachallengingendeavorthatneedstobedonecarefully.Wheredoyou
start? What do you have to consider?
Thisbookwillhelpyoubetterunderstandthedifferentaspectsandchallengesof
evolving development processes. It addresses difficult problems, such as how to
implement processes in highly regulated domains or where to find a suitable
notation for documenting processes. This book emphasizes the need to consider
Software Process Evolution as an important means for catching up with rapid
changes intechnical and marketenvironments. It providesinsights that might help
you manage process evolution. It gives plenty of tips, e.g., how to cope with the
threat of disruption from a process perspective. In addition, it provides many
examples and cases on how to deal with software evolution in practice.
Why a Book on Managing Process Evolution?
Many organizations need to transform their business to the next level. In order to
benefit from leading-edge technologies, catch up with the digital transformation,
and continuously innovate and renew business models, companies have to quickly
1QuotetakenfromatweetfromDavidEvans.
ix
x Preface
adapt and change the ways they develop products and services. As software is the
keytothistransformation,thewaysinwhichmodernsoftwareisdevelopedneedto
change accordingly.
Another important driver for process evolution is the need to mitigate software
risks. Basically, a considerable share of software risk is process-based [3]. For
example, there have been several incidents which could have been avoided with
appropriate coding standards and tools. Although these standards and tools are
widely available, they are either not applied or not appropriately applied in many
situations. Because this is normally caused by the way work and people are
organizedandworkiscarriedoutitisasoftwareprocessissue.Companiesneedto
find ways to ensure that process models are properly defined and, furthermore, are
appropriately applied while not hindering the creativity of, e.g., designers or
developers. To do this effectively, defining and deploying adequate software pro-
cesses usually requires fostering the evolution of existing processes and their
underlying models towards ones that suit better.
Today, there exists a variety of software processes ranging from generic and
domain-specific standards, from agile methods to comprehensive process engi-
neeringframeworks.Sincesoftwareprocessesmaycontainuptohundredsoreven
thousands of elements, the management of a software process is a demanding task
and, therefore, many companies install whole departments dealing with software
process improvement and management. In practice, especially in large organiza-
tions, we can observe some interesting gaps:
(cid:129) Development teamstend toapplyagile methods whilethehosting organization
focuses on “classic” structured development processes [5, 6].
(cid:129) Implemented development processes in projects differ from what has been
defined [4].
(cid:129) Evolving software technologies and platforms require a parallel evolution of
software processes to accommodate the rapid changes. However, this
co-evolution does not appropriately take place.
One main reason for these gaps is different mindsets. For instance, program man-
agers and quality assurance people need planned and directed processes for certi-
fication, budgeting, and compliance business. Developers need flexibility and
processes which support creative work. Business managers need processes that
allow for fast results and flexible feature delivery. Moreover, due to technology
evolution, business evolves. This requires that emerging markets must be addres-
sed, new technologies should be adopted, and globally distributed development
becomes more and more important.
Apart from the big “global players,” process evolution is also highly important
for small and medium-sized companies. Such companies typically neither have
comprehensive process models nor process engineering groups, and often have to
trust in a common understanding of principles and applied practices. However,
theseprinciplesandpracticesneedtobecontinuouslyvalidatedagainsthigherlevel
goals (such as business strategies) and potentially changed in order to secure and
maintain the company’s position in the market place [2]. One example for such a
Preface xi
change is the increasing focus on value-delivery [1]. Regardless of the company
size, a major challenge that companies face is to provide all stakeholders with
flexible processes that:
(cid:129) Are driven by the needs of the different stakeholders,
(cid:129) Have clear links to higher level goals of an organization,
(cid:129) Provide interfaces that are compatible with organizational structures,
(cid:129) Are supported by tools for modeling, enactment, analyses, and evolution,
(cid:129) Can be tailored to individual project goals and characteristics,
(cid:129) Offer adaptability and elasticityto accommodate and supporttechnological and
organizational innovations and evolutions.
This book focuses on the design, development, management, governance, and
applicationof evolving software processesthat arealignedwith changing business
objectives, such as expansion to new domains or moving to global production. In
the context of evolving business, it addresses the complete software process life-
cycle, from initial definition of a product to systematic improvement.
Who Should Read This Book?
This book is aimed at anyone interested in understanding and organizing software
development tasks in an organization. The experiences and ideas in this book are
useful for both those who are unfamiliar with software process improvement and
wanttogetanoverviewofthedifferentaspectsofthetopic,andthoseexpertswith
manyyearsofexperience.Inparticular,thepresentbookaddressesresearchersand
Ph.D. students in the area of Software & Systems Engineering and Information
Systems, who study advanced topics of organizing and managing (software
development) projects and process improvements projects. Furthermore, the book
addresses practitioners, consultants, and coaches involved in software-related
change management and software process improvement projects, and who want to
learn about challenges and state-of-the-art techniques and experiences regarding
their application to problems in different application domains.
How is the Book Organized?
This book is organized in three parts (Fig. 1). Part 1 focuses on software business
transformation,itschallenges,andaddressesthequestionsaboutwhichprocess(es)
touseandadapt,andhowtoorganizeprocessimprovementprograms.InChap.1,
Tony Wasserman discusses short lifecycle projects and how “low-ceremony pro-
cesses” help shorting project iterations. In this context, in Chap. 2, Diebold and
Zehlerdiscussthe“right”degreeofagilityinrichsoftwareprocesses—howtofind
and how to achieve this. The challenge of implementing agile software