Professional Software Development Shorter Schedules, Higher

Professional Software
Shorter Schedules, Higher Quality
Products, More Successful
Projects, Enhanced Careers
by Steve McConnell
Matt Sawka
About the Author
Steve McConnell is has a bachelor's degree Whitman
College and a master’s degree in Software Engineering
from Seattle University.
He has been working in the Software industry since 1984,
including being a consultant for Microsoft.
From 1998-2002, he was Editor and Chief of IEEE Software
He is currently on a panel of experts that advises the
Software Engineering Body of Knowledge project and is a
Vice Chair of the IEEE Professional Practices Committee.
Currently the CEO and Chief Software Engineer of Contrux
What is software engineering?
How does software engineering relate to computer science?
Why isn’t regular computer programming good enough?
Why do we need a profession of software engineering?
Why is engineering the best model for a software development
In what ways do effective practices vary from project to project (or
company to company), and in what ways are they usually the same?
What can organizations do to support a professional approach to
software development?
What can individual software developers do to become full-fledged
What can the software industry as a whole do to create a true
profession of software engineering?
The Software Tarpit
Wrestling with Dinosaurs
Fool’s Gold
Cargo Cult Software
Body of Knowledge
Novum Organum
Individual Professionalism
Orphans Preferred
Raising Your Software
Building the Community
Programmer Writing
Organizational Professionalism
Software Gold Rushes
Business Case for Better Software
Ptolemaic Reasoning
Quantifying Personnel Factors
Construx’s Professional
Development Program
Industry Professionalism
Engineering a Profession
Hard Knocks
Stinking Badges
The Professional’s Code
Part One
The Software Tar Pit
Chapter 1:
Wrestling with Dinosaurs
“He that will not apply new remedies
must expect new evils,
for time is the greatest innovator.”
–Francis Bacon
Wrestling with Dinosaurs
“No scene from prehistory is quite so vivid as that of the moral
struggles of great beasts in the tar pits. In the mind’s eye one sees
dinosaurs, mammoths, and sabertoothed tigers struggling against the
grip of the tar. The fiercer the struggle, the more entangling the tar,
and no beast is so strong and so skillfully but that he ultimately sinks.
Large-system programming has over the past decade been such a tar
pit, and many great and powerful beasts have thrashed violently in it
Most have emerged with running systems – few have met goals,
schedules, and budgets Large and small, massive or wiry, team after
team has become entangled in the tar No one thing seems to cause
the difficulty – any particular paw can be pulled away. But the
accumulation of simultaneous and interacting factors brings slower and
slower motion. Everyone seems to have been surprised by the
stickiness of the problem, and it is hard to discern the nature of it But
we must try to understand it if we are to solve it.
-The Mythical Man-Month (p. 3), Fred Brooks. 1975
Wrestling with Dinosaurs
Has anything changed?
 ~75% of all medium-sized projects and 90% of or more of large
projects are subject to excessive schedule pressure.
 Overtime is the standard.
► “In
many companies, programmers faced with deadlines have been
known to spend nights in the offices.” – Fortune, 1967
 Windows NT was projected to take 1500 staff-years to complete.
► OS/360
took more than 3 times this amount in 1966.
 The advent of web-development poses the question is delivering
software quickly better than delivering quality software? McConnell
argues that this is not a new phenomena.
Chapter 2:
Fool’s Gold
“Hope is a good breakfast,
but it is a bad supper.”
–Francis Bacon
Fool’s Gold
► During
the California Gold Rush of the late 1800s,
many were deceived by fool’s gold – iron pyrite,
which has the same appearance as Gold.
► The last fifty years in software development has
produced similar results. Software Engineers
seeking solid processes and standards have mostly
found fool’s gold, software practices that appear
useless on first glance, but are actually flaky,
brittle, and virtually valueless.
Fool’s Gold
► Problem:
Looking back further in history, how
could we build one of the ancient pyramids from
ancient Egypt?
 The stone blocks need to be moved 10,000 meters from
the river to their final resting place.
 You have 100 days to move the blocks.
 You have 20 people to move the blocks.
 You are allowed to use any method you’d like.
► On
average, you must move the blocks 100
meters a day (or have a method that will speed up
the process later on).
► How would you accomplish this?
Fool’s Gold
► Option
 Immediately start pushing the blocks with brute force.
 Result: You move the block 10 meters a day and fall 90
meters/day behind.
Fool’s Gold
Option 2:
 Analyze the problem. You decide to cut down several trees and
use them as rollers to move the block. Create a smooth, level
roadway to push forward.
 Result: Although there is an initial investment to find the trees, it
still will pay off.
Fool’s Gold
Option 3:
 Implement option 2, but balance the pushers with pullers and add
log-movers, so that a group is always moving extra logs in front of
the block.
 Result: An improvement on option 2, the block will be continuously
Fool’s Gold
► How
does software development relate to ancient
Egyptian pyramids?
 Change the pyramid block to a software development
► You
have 100 days to complete a project. This means you
either have to complete 1/100 of the source code each day or
you need to schedule some parts taking less time than others.
 Avoid “last-minute” syndrome (Option 1)
► The
development team has little concern near the beginning of
the project, but has a frenzied push at the end of the
development cycle.
Fool’s Gold
How does software development relate to ancient Egyptian
 Also avoid “code-and-fix” development (Option 2)
► Put
a brute force of programmers on the project without properly
planning or design of the software. The project team will show initial
progress but will be unable to finish strong. Most effort will go into
Fool’s Gold
► The
Silver Bullet Syndrome
 An elephant could be capture to pull block. However, after it is
captured, it tramples 2 workers and runs off. The managers then
think they should’ve spent time learning to handle the elephant.
Fool’s Gold
As project planning matures, the amount of effort spent
on later stages should be manageable.
► Despite the downfalls, “code-and-fix” is still heavily used
today for two reasons:
 It provides initial signs of progress
 It requires no training
Fool’s Gold
► Focusing
on Quality
 If an organization can remove ~95% of their defects
before a release, they can minimize the amount of
effort spent on correcting them later.
Fool’s Gold
► Software
isn’t “soft”.
► Let’s
suppose you are designing a system to print
a set of 5 reports, and eventually 10 reports.
 How the reports can be “soft”:
► Is
ten an upper limit on thee number of reports?
► Will the future reports be similar to the initial five reports?
► Will all of the reports always be printed?
► Will they always be printed in the same order?
► To what extent will the user be able to customize the reports?
► Will users be allowed to define their own reports?
► Will the reports be customizable and definable on the fly?
► Will the repots be translate to other languages?
Fool’s Gold
 How the reports can be “hard”:
► Defining
more than 10 reports.
► Defining a new report that is different form the initial set of reports.
► Printing a subset of the reports.
► Printing the reports in a user-defined order.
► Allowing the user to customize reports.
► Translating the repots to another language that users the Latin
► Translating the reports to a another language that uses a non-Latin
alphabet or reads right-to-left.
Bottom Line: Flexibility costs money. Limiting flexibility
can save money in the short term, but can incur higher
costs later on the development life-cycle.
Chapter 3:
Cargo Cult Software Engineering
“In the South seas there is a cargo cult of people. During the war they
saw airplanes with lots of good materials, and they want the same
thing to happen now. So they’ve arranged to make things like
runaways, to put fires along the sides of runways to make a wooden
hut for a man to sit in, with two wooden pieces on his head for
headphones and bars of bamboo sticking out like antennas – he’s the
controller – and they wait for the airplanes to land They’re doing
everything right. The form is perfect. It looks exactly the way it
looked before. But it doesn’t work. No airplanes land. So I call these
things cargo cult science, because they follow all the apparent precepts
and forms of scientific investigation, but they’re missing something
essential, because the planes don’t land.”
–Richard Feynman
Cargo Cult Software Engineering
► Process-oriented
vs. Commitment-oriented
Development styles
 Process-oriented
► Relies
on a carefully defined process, planning, scheduling,
and directly application of Software Engineering bestpractices.
► This succeeds because the organization is constantly
improving on their best practices.
 Commitment-oriented
► Also
known as “hero-oriented” or “individual empowerment”,
this style relies on hiring the best people and asking them for
total commitment to a project. They work with completely
autonomy something work 60-100 hours a week until a
project is completed.
► This style succeeds because it utilizes individual motivation.
Cargo Cult Software Engineering
► Imposters:
 Process-oriented imposter (Bureaucratic)
► Observes
successful companies using best-practices (such as
NASA’s Software Engineering Laboratory), see that they have
many meetings and documents and emulate the deliverables
 Commitment-oriented imposter
► Observe
successful companies like Microsoft, and emphasize
the long hours and large compensation packages, not the fact
that the employees of Microsoft love to create software.
► Cargo-cult
engineering is simply using a
development style “just because” or “because
the company requires it”, rather than using the
style advantageously.
Chapter 4:
Software Engineering, Not Computer
“A scientist builds in order to learn; an
engineer learns in order to build.”
–Fred Brooks
SE not CS
Like other engineering disciplines, Software Engineering
can be tailored for several objectives:
Minimal defects
Maximum user satisfaction
Minimal response time
Good maintainability
Good extendibility
High robustness
High correctness.
Short schedule
Predictable delivery date
Low cost
Small team size
Flexibility to make mid-project feature-set changes.
Unlike other disciplines in which risk relies on the physical
materials, Software Engineering carries risk in optimizing
the project goals.
SE not CS
► What
is the best way to think of software
development? Is it a science? Art? Craft?
Something else?
 ~40% of developers today have degrees in Computer
 Scientists learn what is true, how to test hypotheses,
and how to extend their knowledge.
 Engineers learn what is true, useful, and how to apply
their knowledge to solve a problem.
► Is
Software Engineering just a buzzword?
 Some object to Software development being classified
as engineering because the commercial market
doesn’t allow for the careful, time-consuming
engineering process to be completed.
Chapter 5:
Body of Knowledge
“Truth will sooner come out of error than from
–Francis Bacon
Body of Knowledge
► To
be an expert in a field, a person needs to
know around 50,000 pieces of information.
► But with software engineering evolving, how can
anyone know enough to be considered an
► Essence vs. Accident
 No Silver Bullet – Essence and Accidents of Software
Engineering – Fred Brooks, 1987.
 Essence – properties that something must have.
► Wheels
on a car
► Software engineering: Specification, Design, and Verification.
 Accident – optional properties.
► Air-Conditioning
► Software
Engineering: Coding and testing.
Body of Knowledge
► Computer
programs are complex.
 At the center, the goal is to define underlying realworld concepts and debugging that understanding.
► Software
must be flexible and have the ability to
 If a program is successful, more people will use it and
it will be adapted to be used outside of its original
► Software
is “invisible.”
 It’s not possible to create a 2-d or 3-d geometric
model of the system.
Body of Knowledge
► In
1968, NATO held its
first conference on
Software Engineering.
► McConnell estimates
that in 1968, the
average half-life of
knowledge was about
10 years, with only
about 20% of that
knowledge at the
Stable Core.
Body of Knowledge
► By
2003, McConnell
estimates that the
average half-life of
knowledge was
about 30 years,
with about 50% of
that knowledge at
the Stable Core.
Body of Knowledge
► How
can we categorize the Software
Engineering “body of knowledge” (SWEBOK)?
Body of Knowledge
► What
sources contribute to the SWEBOK?
Body of Knowledge
► What
information is contained within SWEBOK?
► SWEBOK Knowledge Areas
 Software Requirements
► The
discovery, documentation, and analysis of the functions
to be implemented in software.
 Software Design
► Definition
of the basic structure of the system at the
architectural and detailed levels, division into modules,
definition of interfaces for modules, and choice of algorithms
within modules.
 Software Construction
► Implementation
of the software including detailed design,
coding, debugging, unit testing, technical reviews, and
performance optimization. This area overlaps somewhat with
Software Design and Software Testing.
Body of Knowledge
Knowledge Areas
 Software Testing
► All
activities associated with executing software to detect
defects and evaluate features. This includes planning, testcase design as well as the tests themselves: development,
unit, component, integration, system, regression, stress, and
 Software Maintenance
► Revision
and enhancement of existing software, related
documentation, and tests.
 Software Configuration Management
► Identification,
documentation, change control of all
deliverables generated on a project (source code, content,
requirements, etc…).
 Software Quality
► All
activities associated with providing confidence that a
software item conforms or will conform to technical
Body of Knowledge
Knowledge Areas
 Software Engineering Management Plan
tracking and controlling of software
projects, work, and organizations.
 Software Engineering Tools and Methods
and methodologies support (CASE tools,
reusable libraries, formal methods and practices,
 Software Engineering Process
related to improving development quality,
timeliness, productivity, and other characteristics.
Chapter 6:
Novum Organum
“A prudent question is one-half of wisdom.”
–Francis Bacon
Novum Organum
In the 1620s, Francis Bacon
published Instauratio Magna
which attempted to redefine
scientific inquiry. Within this
work is an essay, Novum
Organum which challenges his
colleagues to focus on scientific
methodologies rather than
deductive reasoning when
studying the world. His view of
the scientific method had three
 Purge your mind of prejudices
 Collect observations and experiences
 Stop, survey what you’ve seen, and
draw an initial conclusion.
Novum Organum
► What
does it mean to have a Software
Engineering “profession”?
 According to the Code of Federal Regulations, a
profession has:
requirement for extensive learning and training
► A code of ethics imposing standards higher than those normally
tolerated in the marketplace.
► A disciplinary system for professionals who breach the code
► A primary emphasis on social responsibility over strictly
individual gain, and a corresponding duty of its members to
behave as members f a disciplined and honorable profession.
► A prerequisite of a license priori to admission to practice.
Novum Organum
► Is
Engineering a
 The Software
Engineering Institute
(SEI) has identified 8
elements of a mature
Novum Organum
► Elements
to a mature profession:
 Initial Professional Education
► Completing
a university program in their field.
 Accreditation
► The
university program is accredited by an oversight body that
determines if the program provides adequate education.
Accreditation Board for Engineering and Technology (ABET)
oversees American engineering programs.
 Skills Development
► Education
is not enough to develop full professional capabilities,
some type of further experience is needed to perform the job
individually with a satisfactory level of competence.
Novum Organum
► Elements
to a mature profession:
 Certification
professional is requirements to pass one or more exams that
ensures they have a minimum level of knowledge.
 Licensing
► Similar
to certification, this is a mandatory exam administered by
a overrunning authority.
 Professional Development
► Ongoing
professional education to maintain or improve a worker’s
knowledge and skills.
Novum Organum
► Elements
to a mature profession:
 Professional Societies
Community of like-minded individuals who put their professional
standards above their self-interest.
 Code of Ethics
► Ensures
that a profession’s practitioners behave responsibly.
 *Organizational Certification
► Organizations
must also be certified.
Novum Organum
► Maturity
Levels for the elements
 Nonexistence
► The
element simply doesn’t exist.
 Ad Hoc
► The
element exists, but only in isolated instances
 Established
► The
element exists and is clearly identifiable.
 Maturing
► The
element has existed for many years and is being maintained
and improved.
Novum Organum
► How
does Software Engineering rank?
Initial Professional Education – Ad Hoc to Established
Accreditation – Established
Skills Development – Established
Licensing – Ad Hoc
Professional Development – Ad Hoc
Professional Societies – Established
Code of Ethics – Established
Organizational Certification - Established
The Software Tar Pit Summary
Wrestling with Dinosaurs
 Has anything changed?
► Fool’s Gold
 Building the Egyptian Pyramids
 Code-And-Fix, Last-Minute, Silver-Bullet
 Software isn’t “soft.”
► Cargo Cult Software Engineering
 Process-Oriented vs. Commitment-oriented
► Software Engineering Not Computer Science
► Body of Knowledge
► Novum Organum
 What is a profession?
Part Two
Individual Professionalism
Chapter 7:
Orphans Preferred
“Wanted: Young, skinny, wirey fellows not over 18. Must be
expert riders willing to risk death daily. Orphans preferred.
Wages $25/week.”
-Pony Express, 1860
“We realize the skills, intellect, and personality we seek are
rare, and our compensation plan reflects that. In return,
success – overcoming all obstacles to create applications
on time and within budget.”
–Jobs Rated Almanac, 1995 for an SE posting.
Orphans Preferred
“The stereotypical programmer is a shy young man who works in a
darkened room, intensely concentrating on magical incantations that
make the computer do his bidding. He can concentrate for 12 to 16
hours at a time, often working through the night to make his artistic
vision a reality. He subsists on pizza and Twinkies. When interrupted,
the programming creature responds violently, hurling strings of cryptic
acronyms at his interrupter – “TCP/IP, RPC, RCS, ACM, and IEEE!” he
yells. The programmer breaks his intense concentration only to attend
Star Trek conventions and watch Monty Python reruns. He is sometimes
regarded as an indispensable genius, sometimes as an eccentric artist.
Vital information is stored in his head and his head alone. He is secure
in his job, knowing that, valuable as he is, precious few people compete
for his job.”
Orphans Preferred
Meyers-Briggs Type Indicator
 Extroversion (E) vs. Introversion (I)
► Extroverts
are focused on the outside world, people. Introverts
focus on the world of ideas.
 Sensing (S) vs. Intuition (N)
► How
a person deals with decision-making data. Sensing persons
focus on facts, concrete data and experience. Intuitive people look
for possibilities and focus on conceptual theories.
 Thinking (T) vs. Feeling (F)
► How
a person makes a decision. The thinkers make objective,
analytic decisions, where as feelers rely on emotions and feelings.
 Perceiving (P) vs. Judging (J)
► The
perceiving person is flexible and likes open-ended possibilities,
where as the judging person prefers control and order.
Orphans Preferred
► Meyers-Briggs
and Software Developers
 Most Software Developers (25-40%) are ISTJ.
► 50-75%
of programmers are introverts, compared to 25% in
the general population.
 About 60% of software developers have at least a Bachelors,
compared to 30% of the general population.
► 80-90%
of programmers are Thinking (T) compared to 50%
of the general population.
► There’s a 50-50% split of programmers between Sensing (S)
and Intuition (N).
Orphans Preferred
► Meyers-Briggs
and Designers
 Great Designers…
can general move in-between categories.
► have a mastery of common tools.
► aren’t afraid of complexity.
► seek out constructive criticism.
► have experienced failed projects.
► are not afraid of the brute-force approach.
► must be creative.
► have a restless desire to create.
Orphans Preferred
► Total
and Absolute Commitment
 Although working 12 to 16 hours may seem extreme,
it’s not out of the realm of possibility. Many Microsoft
programmers working this or more during the
development of Windows NT.
► “Work
pervades their existence. Friends fade into the
background. The ties of marriage fray or rip apart. Children
are neglected or deferred. Hobbies wither. Computer code
comes to mean everything. If private dreams are nursed at
al, it is only to ease the pain of creating NT.” – P. Zachary
 Programmers tend to show a loyalty to the project,
even if their loyalty to the company is waning.
 Some programmers aim to be “hero” programmers,
who take on mountains of work and hours.
Orphans Preferred
 The average programmer
age peaks between 30-35
years old (which is about
10 years younger than
most professions).
 72% of computer and
information science BSs
and 83% of Ph. D’s were
 Only 17% of high-schoolers
taking the AP Computer
Science test were female.
This is the lowest % of all
AP tests.
Highest level of
% of
High school or less
Some college,
no degree
Associate’s degree
Bachelor’s degree
Graduate degree
Orphans Preferred
► Job
Job breakdown for software workers
Current Software prsnl in US
Computer and information scientists, research
Computer programmers
Computer software engineers, applications
Computer software engineers, systems software
Computer systems analysts
Network systems and data comm. analysts
Other computer specialists
Chapter 8:
Raising Your Software Consciousness
“If a man will begin with certainties, he shall
end in doubts; but if he will be content to
begin with doubts, he shall end in
-Francis Bacon
Raising Your Software Consciousness
the 1970s, Charles Reich published, The
Greening of America, which identified 3 types of
► In
 Consciousness I (Con I): Pioneer mentality
► Great
emphasis on independence and self-satisfaction
 Consciousness II (Con II): Gray flannel suit mentality
► The
corporate man. Knowing how to get along with other and
playing by the rules
 Consciousness III (Con III): Enlightened Independence
► Operates
on the basis of principles, with little regard for the
rules of Con II and without the selfishness of Con I.
Raising Your Software Consciousness
the 1970s, Charles Reich published, The
Greening of America, which identified 3 types of
► In
 Consciousness I (Con I): Pioneer mentality
► Great
emphasis on independence and self-satisfaction
 Consciousness II (Con II): Gray flannel suit mentality
► The
corporate man. Knowing how to get along with other and
playing by the rules
 Consciousness III (Con III): Enlightened Independence
► Operates
on the basis of principles, with little regard for the
rules of Con II and without the selfishness of Con I.
Raising Your Software Consciousness
► Consciousness
and Software Development
 Con I – Self Reliance
► Software
developers who are “Lone Rangers”. They have little
tolerance for other ideas.
► Little training is needed. This approach works adequately in
small projects.
 Con II – Rules
► Rules
allow programmers to work with others.
 Con III – Principles
► Programmers
understand that the rules from Con II are based
on principle. Programmers focus on the underlying effective of
their actions on software development.
Chapter 9:
Building the Community
“If any human being earnestly desires to push on
the new discoveries instead of just retaining and
using the old; to win victories over Nature as a
worker rather than over hostile critics as a
disputant; to attain, in fact, clear and
demonstrative knowledge instead of attractive and
probable theory; we invite him as a true son of
Science to join our ranks.”
-Francis Bacon
Building the Community
► It’s
important to remember that we’re not simply
lone-ranger programmers, or even programmers
for one company, but rather a community of
► These
organizations can provide solutions to
common problems or articles to further your
understanding of the field.
Chapter 10:
Architects and Carpenters
“Engineers produce plans Builders implement
the plans to produce a product.”
-Terri Maginnis
Architects and Carpenters
As Software Engineering continues to develop, a wider
range of abilities will begin to distinguish itself.
 Average software developers
 Highly skilled software developers
 Unlicensed software engineers/certified software technologists
 Professional Software Engineers
The average person who earns a professional degree earns
50% more than someone without.
► The average person who earns a master’s degree will earn
25% more than someone without.
Architects and Carpenters
► Job
 The Surgical Team,
proposed by Fred
Brooks in The
Mythical ManMonth
Architects and Carpenters
► Job
specializations today
 Technology
► Software
 Knowledge of specific technologies (Microsoft, Novell, Oracle,
 Software Engineering
► Software
 Architecture, configuration control, cost estimation, customer
support, database administration, education and training, function
point counting, human factors, information systems, integration,
maintenance and enhancement, measurement, network, package
acquisition, performance, planning, process improvement, quality
assurance, requirements, reusability, standards, systems software
support, technical writing, testing, tools development.
Architects and Carpenters
► Team
Construction lead
Design lead
Planning and tracking lead
Project business manager
Quality assurance lead
Requirements lead
Chapter 11:
Programmer Writing
“Read not to contradict and confute, nor to
believe and tae for granted, nor to find talk
and discourse, but to weigh and consider.”
-Francis Bacon
Programmer Writing
“The gap between the best software engineering practice
and the average practice is very wide – perhaps wider than
in any other engineering discipline. A tool that
disseminates good practice would be important.”
– No Silver Bullet, Fred Brooks
► According to McConnell, there are six types of authors:
Recent retirees
University professors
Seminar instructors
Think-tank developers
Developers working on production software.
Who should be writing our documentation?
Programmer Writing
“In this distribution of functions, the scholar is the delegated intellect.
In the right state, he is, Man Thinking. In the degenerate state, when
the victim of society, he tends to become a mere thinker, or, still worse,
the parrot of other men's thinking.
…I learn immediately from any speaker how much he has already
lived, through the poverty or the splendor of his speech. Life lies
behind us as the quarry from whence we get tiles and copestones for
the masonry of today. This is the way to learn grammar. Colleges and
books only copy the language which the field and the work-yard
-Excepts from ’The American Scholar’ by Ralph Waldo
Emerson (1837),
Is there a disconnect between academia and the work-force? Are we
still thinking for ourselves? Or merely “thinking” of what others have
thought through?
Programmer Writing
“James Fenimore Cooper syndrome”
 In The Deerslayer, Cooper writes that 6 Indians climbed onto
sapling to board a scow coming downstream.
 Mark Twain described this as a fantasy situation. Because of his
knowledge of riverboat piloting, he knew this situation was not
plausible with the sapling or with the dimensions of the scow.
 This syndrome represents a practitioner calling into the question
the writings of a scholar.
Does the Software Engineering literature suffer from James
Fenimore Cooper Syndrome?
Individual Professionalism
Orphans Preferred
 Meyers Briggs and Software Engineering demographics
Raising your Software Consciousness
 The Greening of America
Building the Community
 IEEE and ACM
Architects and Carpenters
 Specializations
Programmer Writing
 ‘The American Scholar’
 James Fenimore Cooper Syndrome
Part Three
Organizational Professionalism
Chapter 12:
Software Gold Rushes
“Prosperity doth best discover vice, but
adversity doth best discover virtue.”
-Francis Bacon
“The root of all superstition is that men
observe when a thing hits,
but not when it misses.”
-Francis Bacon
Software Gold Rushes
In 1848, gold was discovered in riverbeds in California.
Many self-proclaimed entrepreneurs set out to make their
fortune. This made up the California Gold Rush. But, by
mid-1849, the majority of the easily found gold was
collected. Many miners would spent hours a day digging
through freezing water looking for any traces of the metal.
By the 1850s, most miners had joined corporations to
continue their hunts.
Software development experienced a similar phenomena.
There are a few success stories (Bill Gates and Paul Allen of
Microsoft; Steve Jobs and Steve Wozniak of Apple, Bob
Frankston and Dan Bricklin of VisiCalc), but many failures as
Software Gold Rushes
Much of the “software gold rush” is characterized by highrisk projects, long hours, hacking, informal or no processes,
little-to-no documentation, and very little quality assurance.
Post-gold rush development has en emphasis on lower-risk,
high capital projects, larger teams, formal processes, and
general standards. There isn’t an effort to rush projects out
the door, but rather do thorough testing.
Post-gold rush consumers are generally more demanding on
the products that they are looking to buy.
Many companies that survive a gold-rush would not be able
to survive a second.
Chapter 13:
Business Case for Better Software Practices
“When you can measure what you are
speaking about, and express it in numbers,
you know something about it; but when you
cannot measure it, when you cannot
express it in numbers, your knowledge is of
a meager and unsatisfactory kind.”
-Lord Kelvin, 1893
Business Case for Better Software Practices
► Development
practices pay off
 In 1994 James Herbsleb prepared a study of 13
organizations’ “business value” (similar to Return On
► In
1995, systemic improvements increased the ROI anywhere
from 500-900%.
 In 1997, Rini van Solingen found an increase ranging
from 700-1900%.
 In 2000, Caspers Jones found that the ROI could easily
be in the double digits (over 1000%).
 In 2001, Watts Humphrey found an ROI increase of
Business Case for Better Software Practices
► Current
organizational effectiveness
 A study funded by the SEI found that those organizations who implemented
a change to their development practice found an average of 35% gain in
productivity, 19% schedule reduction, and post-release defects were reduced
by 39%.
 The same study found that for the best-case scenarios, 58% productivity
could be gained over 4 years, a compounded gain of 500% was achieved, a
23%/year schedule reduction (with a 91% reduction over 6 years), and postrelease defects were reduced by 99%.
Business Case for Better Software Practices
BDN International
ROI 300%
Boeing Information System
Estimates within ~20%, $5.5 million saved
in 1 year, ROI 775%
Computer Sciences Corporation
65% reduction in error rate
General Dynamics Decision Systems
70% reduction in rework; 94% defect rate
reduction (drr), 2.9*productivity gain
Harris ISD DPL
90% drr, 2.5*productivity gain, ROI 900%
$2 million reduction in costs, ROI 500%
IBM Toronto
90% drr, 80% reduction in rework
Motorola GED
2-3*productivity gain, 2-7*cycle reduction
time, ROI 677%
ROI 750%
ROI 770%
4*reduction in released defects
Business Case for Better Software Practices
90% reduction in released defects
Defect 1/10 industry aver, customer
satisfaction up from 60% to 91%
Texas Instruments – Systems
90% reduction in delivered defects
Thomson CSF
ROI 360%
U.S. Navy
ROI 410%
USAF Ogden Air Logistics Center
ROI 1,900%
USAF Oklahoma City Air Logistics
ROI 635%
USAF Tinker Air Force Base
ROI 600%
Business Case for Better Software Practices
12-month ROI
36-month ROI
Formal code inspections
Formal design inspections
Long-range technology planning
Cost and quality estimation tools
Productivity measurements
Process assessments
Management training
Technical staff training
Business Case for Better Software Practices
► Performance
improvements with the Capability
Maturity Model (CMM)
Business Case for Better Software Practices
Performance improvements with the Capability Maturity Model (CMM)
 In general for average organizations,
Effort = 2.94 * (KLOC)^1.10
 NASA’s Software Engineering Laboratory’s (SEL) effort calculation has
Effort = 1.27 * (KLOC)^.986
 The .986 is especially important because it means they are achieving
a slight economy of scale.
 Only three of the 22 factors used to calculate effort were changeable
at the individual project management level: Level of Documentation,
Architecture and Risk Resolution, and Development for Reuse. Most
of these factors are at the organizational level and cannot be easily
Business Case for Better Software Practices
10 questions to ask about software activities
How much are you spending on software development?
What percentage of your projects are currently on time and on budget?
What is the average schedule and budget overrun for your projects?
Which of your current projects are most likely to fail outright?
What percentage of your project costs arises from avoidable rework?
How satisfied (quantitatively) are users of your software?
How do the skills of your staff compare to the industry averages?
How do the capabilities of your organization compare to similar
How much (quantitatively) has your productivity improved in the past
12 months?
What is your plan for improving the skills of your staff and the
effectiveness of your organization?
Chapter 14:
Ptolemaic Reasoning
“All models are wrong;
some models are useful.”
-George Box
“Knowledge itself is power.”
-Francis Bacon
Ptolemaic Reasoning
► Ptolemy
was an astronomer who lived around
A.D. 100 and theorized that the sun revolved
around the earth.
► His theory was eventually replaced by Copernicus
when his data showed findings that contradicted
► In the same way, real-world data supports an
object-oriented development process.
Ptolemaic Reasoning
► Capability
Maturity Model
 Developed by the Software Engineering Institute.
 There are five software levels:
► Level
1: Initial
 Software development is chaotic. This is the default level where
the organization generally relies on the “code-and-fix
development” model.
► Level
2: Repeatable
 Basic project management practices are established on a
project-by-project basis. The strength of the organization lies
on its ability to repeated experiences.
Ptolemaic Reasoning
► Capability
► Level
Maturity Model
3: Defined
 The organization adopts standardized technical and
management processes across the organization. A group is
assigned to monitor the software process.
► Level
4: Managed
 Project outcomes become highly predictable. A standard
process is identified and variations can be detected.
► Level
5: Optimizing
 The focus of the organization is on proactive identification of
process improvements. It has the ability to measure changes to
the process and identify their strengths and weaknesses.
 Conway’s Law: The structure of a computer program
reflects the structure of the organization that built it.
Ptolemaic Reasoning
► Navigating
the Capability Maturity Model
Ptolemaic Reasoning
► Navigating
the Capability Maturity Model
Ptolemaic Reasoning
CMM and Risk Management
 Cheyenne Mountain ATAMS project
► The
project team was able to…
 complete the project in 1/5th of the projected time.
 complete the project in ½ of the estimated time.
► Overall,
they were able to deliver the software one month ahead of
schedule and within budget.
► Eighteen months after release, only two defects were discovered.
Would developers like CMM?
 In a survey of 50 organizations, 20% of level 1 organization rated
staff morale as “good” or “excellent”.
 At Level 2, 50% of the staff rated morale as “good” or
 At Level 3, 60% of the staff rated morale as “good” or
 Most developers working in a level 5 environment don’t want to
work for an organization that is less than level 5.
Ptolemaic Reasoning
► What
does it take to implement CMM?
 Commitment from top management (and followthrough)
 Establishment of a Software Engineering process
group (sometimes multiple groups).
 Appropriate training in middle management and
various technical positions.
► What
is success?
 Success = Planning * Execution
 As an organization moves up CMM levels, they will find
that lower levels seem valuable and less effective.
Chapter 15:
Quantifying Personnel Factors
“Personnel attributes and human relations
activities provide by far the largest source of
opportunity for improving software
-Barry W. Boehm
Quantifying Personnel Factors
Personnel Experience
 In a study conducted by Sackman, Eriskon, and Grant, they
found that the variance in time-to-completion for a programming
problem varied almost 20-to-1 among a programmers with 7+
years of experience.
 Other studies have found similar findings ranging anywhere from
5-to-1 to 10-to-1.
 In some cases, the programmers were unable to complete the
 In a group of 7 random programmers, there’s a 50/50 change
that at least one of the programmers will produce negative
 There are 7 factors that are influenced by experience:
Application experience (1.51), Communications factors (1.53),
Language and tool experience (1.43), Personnel continuity
(1.51), Platform experience (1.40), Programmer capability (1.76),
Analyst capability (2.00).
Quantifying Personnel Factors
Physical Environment
 In these war-games, the environment played a factor. Of the top
25% of programmers, most have bigger, private offices (~78 ft2)
with fewer interruptions
 Microsoft provides each group with a “morale budget” which the
team can spend on anything from plaques, to t-shirts, to dinner
and movies.
 Microsoft allows programmers to accommodate family situations.
 Because programming experience plays such a large factor in
development, many companies have found success in have senior
personnel help with each project.
Chapter 16:
Construx’s Professional Development Program
Construx’s Professional Development Program
► Construx’s
Software Development objectives
 Skills enhancement
► Improve
the skills of the employees
 Career pathing
► Provides
a structured path for improvement and career choice.
 Support for common software job titles
► Have
a full list of titles: software developers, testers, business
analysts, project managers, architects, etc…
 Consistency
► Provide
a consistent means for evaluation
 Generlizability beyond Contrux
► Provide
a generic framework for other companies to model.
Construx’s Professional Development Program
► Construx’s
Knowledge Areas (SWEBOK)
Configuration Management
Engineering Management
Engineering Tools and Methods
Engineering Process
Construx’s Professional Development Program
► Construx’s
Capability Levels
 Introductory
► Employee
can perform basic functions in an area under
supervision. Professional development is occurring.
 Competency
► Employ
can perform effective, independent work and serve as a
role model for the less experienced. Occasionally mentors
 Leadership
► Employee
performs exemplary work and regularly coaches
employees and provides some leadership direction.
 Mastery
► Employee
can perform reference work and has deep experience.
Usually, the employee has taught classes or written papers.
They are recognized outside of Construx.
Construx’s Professional Development Program
► Construx’s
Capability Levels
Introductory Competency Leadership
Introductory Competency Competency
Competency Competency Competency
Competency Competency Leadership
Construx’s Professional Development Program
► Construx’s
Professional Development Ladder
 Each engineer is giving a rating between 9-15, based on
their knowledge and experience. A 12 is considered to
by a full professional.
 Most engineers don’t go beyond because it requires
career-long commitments to the company to the
Software Engineering field.
► Construx’s
 Professional Development Plan, Mentoring Program,
Professional Development Plaques, Training Program,
Salary Structure, SE Discussion groups, Level-12
Organizational Professionalism
Software Gold Rushes
 Gold Rush vs. Post-Gold Rush software
Business Case for Better Software Practices
 Process improvements and their effect on business
Ptolemaic Reasoning
Quantifying Personnel Factors
 Environment, Motivation, etc…
Contrux’s Professional Development Program
Development objectives
Capability Levels
Ladder-based Careers
Part Four
Industry Professionalism
Chapter 17:
Engineering a Profession
“Engineering can provide a life of genuine
satisfaction in many ways, especially
through ministering in a practical manner to
the needs and welfare of mankind.”
-Vannevar Bush
Engineering a Profession
► Engineering
feats of the past
Reims Cathedral
Sydney Opera House
Engineering a Profession
Maturation Cycle of an Engineering Discipline
 There are 3 stages in the development of a discipline
► Craft
 Work is performed by talented amateurs. There is little-to-no concept of
► Commercial
 Workers are more economically driven. The focus is on quality in the
products and production processes.
► Professional
 A corresponding science is developed to better understand the engineering
problem and this science is then applied on a wider scale to many of the
common engineering problems of the discipline.
 As a discipline matures, solutions to common problems are codified
(equations, models, components) so that they can be reused.
Engineering a Profession
The science behind software
 Unlike Physics, the science behind the Civil Engineering that built
the Reims Cathedral, Software Engineering doesn’t have an exact
science behind it.
 Software Engineering artifacts
► Architectures, software
► Design patterns
► Requirements
design procedures
► User Interface elements and procedures
► Estimates
► Planning data (project plans, procedures)
► Test plans, cases, data, procedures
► Technical review procedures
► Source code, construction procedures, integration
► Software configuration management procedures
► Post-project reports
► Organizational structures
Chapter 18:
Hard Knocks
“Natural abilities like natural plants, that need
pruning by study; and studies themselves
do give forth directions too much at large,
except they be bounded in by experience.”
-Francis Bacon
Hard Knocks
Hard Knocks
► Is
Computer Science relevant to Software
 Notice the decline of Computer Science and IS degrees
from 1985-1997.
► The
decline was due to Computer Science curriculum's becoming
irrelevant to the workplace.
 There’s an upsurge from 1998-2000.
► The
Gold Rush and dot-com booms fueled a resurgence
interested for incoming freshmen.
Hard Knocks
► Should
Software Engineering follow the same path
as other engineering disciplines?
Hard Knocks
Academic development of Software Engineering
 The first Master’s degree in Software Engineering was offered by
Seattle University in 1982.
 University of Sheffield’s Computer Science department (UK) offered
an undergraduate Software Engineering degree in 1988.
 Rochester Institute of Technology offered the first US
undergraduate course in 1996.
 25% of Software Engineering programs are offered in the US.
 In 2003, ~24 institutions offered an undergraduate Software
Engineering degree.
 Should Software Engineering focus on software or engineering?
 The Computer Science accrediting board requires that their students
make a ”scholarly contribution” to computer science, but no industry
 ABET requires “non-academic engineering experience.”
Chapter 19:
Stinking Badges
“Badges? We ain’t got no badges. We don’t
need no badges. I don’t have to show you
any stinking badges.”
-Gold Hat Bandito,
The Treasure of the Sierra Madre
Stinking Badges
► Certification
 A voluntary process culminating in an exam that is
administered by a professional society.
 Typically, both education and experience are required.
 There are some certifications that are available.
► Institute
for Certification of Computing Professionals’ Associate
Computing Professional and Certified Computing Professional
► Microsoft’s Microsoft Certified Professional
► Novel’s Certified Network Engineer
► Oracle’s Oracle Certified Professional
► Apple’s Apple Certified Server Engineer
► IEEE’s Certified Software Development Professional.
Stinking Badges
 A mandatory legal process culminating in an exam that is
administered by a governing agency.
 Typically, this is required to protect the public (doctors, architects,
Can Software Engineering be licensed?
 ‘There is no generally agreed upon body of knowledge for software
 ‘Knowledge is software engineering changes so quickly that exams
will be out of date by the time they’re offered.’
► Review
the contents of SWEBOK.
Stinking Badges
Can Software Engineering be licensed?
 ‘No reasonable test for software engineering skill could be put into a
multiple-choice format. Indeed, no exam-based practices could
adequately ensure competency of software engineers.’
► Other
complex disciplines already have exams (medicine, law, etc…).
Why is this any different?
 ‘The breadth of sub-disciplines involved in software would make
licensing all of the sub-disciplines impractical.’
► The
medical profession has several exams for different specialties. Only
software development that will impact the public’s health and/or
welfare would need to be explicitly licensed.
 ‘The fundamentals of Engineering Exam required for licensing
existing professional engineers is inappropriate for those receiving a
computer science degree.’
► Some
undergraduate degree programs focus on the engineering
discipline, but more work would need to be done to adopt this exam.
Stinking Badges
Is licensing a bad idea?
 ‘Licenses would unduly restrict the number of people who could
practice software engineering at a time when demand for software
engineers is increasing.’
► Not
all Software Engineers would be required to have a license.
 ‘When an engineer receives a license, it will be good for life, which
is inappropriate considering that software engineering’s body of
knowledge is changing so rapidly.’
► Review
SWEBOK and note that other licenses help keep professionals
 ‘Licensing can’t guarantee that every individual who’s licensed will
actually be competent. Licensing will give the public a false sense
of security.’
► Like
other disciplines, only certain Software Engineers would need to be
Stinking Badges
More on Licensing
 Texas currently offers a “bootstrap” Professional Engineering license
for Software Engineers.
 One side-effect of being licensed is that you’re legally accountable
for your work.
 Licensed engineers would have the final say on methodology
 3 paths to general licensing
► Specialty
in engineering that focus on software
► Create a software engineering specialty exam
► Create a Professional Software Engineering credential.
Engineering students in Canada receive an iron ring to
symbolize their commitment to society.
Chapter 20:
The Professional’s Code
“Although there is a little bit of Peter Pan in
each of us, maturity brings with it the desire
to contribute to the communal welfare. The
fulfillment of this yearning, I repeat,
provides the engineer with his primary
existential pleasure.”
-Samuel C. Florman
The Professional’s Code
code of ethics is required for all professions.
► Software Engineering Code of Ethics and
Professional Practice (joint venture with ACM/IEEE)
 There are two guiding philosophies…
► “Software
engineers shall commit themselves to making the
analysis, specification, design, development, testing and
maintenance of software a beneficial and respected profession.
In accordance with their commitment to the health, safety and
welfare of the public, software engineers shall adhere to the
following Eight Principles: “ (Preamble short-version)
The Professional’s Code
Software Engineering Code of Ethics and
Professional Practice
There are eight guiding principles…
1. Public: Software engineers shall act consistently with the
public interest.
2. Client and Employer: Software engineers shall act in a manner
that is in the best interests of their client and employer,
consistent with the public interest.
3. Product: Software engineers shall ensure that their products
and related modifications meet the highest professional
standards possible.
4. Judgment: Software engineers shall maintain integrity and
independence in their professional judgment.
The Professional’s Code
Software Engineering Code of Ethics and
Professional Practice
There are eight guiding principles…
5. Management: Software engineering managers and leaders
shall subscribe to and promote an ethical approach to the
management of software development and maintenance .
6. Profession: Software engineers shall advance the integrity and
reputation of the profession consistent with the public interest.
7. Colleagues: Software engineers shall be fair to and supportive
of their colleagues.
8. Self: Software engineers shall participate in lifelong learning
regarding the practice of their profession and shall promote an
ethical approach to the practice of the profession.
The Professional’s Code
Why is the Code needed?
All professions are required to have a Code.
Death-march projects
Low-ball bidding
Developers should provide realistic estimates.
Code-and-fix development
Developers debate unachievable schedules.
Inconsistent with developing quality code.
Knowledge stagnation
Developers cannot perform professionally without keeping upto-date on professional issues.
Chapter 22:
“Q: What are the most exciting/promising software
engineering ideas or techniques on the horizon?
A: I don’t think that the most promising ideas are
on the horizen. They are already here and have
been here for years but are not being used
-David L. Parnas
Best Practices
Year Described
Automated estimation tools
Evolutionary delivery
Productivity environments
Risk management planing
Change board
Throwaway user interface prototyping
JAD sessions
Information hiding
Design for change
Source code control
Incremental integration
Branch-coverage testing
Software Engineering Process Groups
If so many of the best-practices have been in existing for
15+ years, why haven’t they been implemented on a
wide-spread scale?
Many organizations fear changing what’s worked previously.
Not all best-practices work in practice.
Many companies fear the risk involved in changing.
How can we diffuse information?
The Agricultural Extension
Research Subsystem
State Extension Agents
Connect the Research Subsystem to the County agents.
County Extension Agents
Research professors producing innovative results that are later
diffused to the group.
Persons who help local farmers choose what innovations that
should use.
SEI acts as software’s Research Subsystem
NASA’s SEL producing guidebooks ad training courses.
Industry Professionalism
Engineering a Profession
 What is a profession?
 Is Software Engineering a profession?
Hard Knocks
 History of Software Engineering
Stinking Badges
 Certification vs. Licensing
The Professional’s Code
 Software Engineering Code of Ethics and Professional Practice
 Best-practices
 Diffusing information
The Software Tarpit
Wrestling with Dinosaurs
Fool’s Gold
Cargo Cult Software
Body of Knowledge
Novum Organum
Individual Professionalism
Orphans Preferred
Raising Your Software
Building the Community
Programmer Writing
Organizational Professionalism
Software Gold Rushes
Business Case for Better Software
Ptolemaic Reasoning
Quantifying Personnel Factors
Construx’s Professional
Development Program
Industry Professionalism
Engineering a Profession
Hard Knocks
Stinking Badges
The Professional’s Code