deliberate.
From the electronics books by Forrest M. Mims III to the
back-of-the-napkin design documents favored by agile
developers, hand-drawn figures have certain unique prop-
erties, and we’ll see why a bit later in the book.
1.5 Grateful Acknowledgments
Very special thanks to Ellie Hunt for introducing me to the Drey-
fus model and related nursing research, suffering through my dis-
jointed and rambling prose, keeping me on track, and keeping our
domestic enterprises running like a well-oiled machine. A regular
editor’s job is often difficult and thankless, and mere appreciation
in a preface really doesn’t do it justice. To be editor, mom, and
business manager all at once truly takes skill and patience.
Thanks to my friends on the Pragmatic Wetware mailing list
and reviewers, including Bert Bates, Don Gray, Ron Green,
Shawn Hartstock, Dierk Koenig, Niclas Nilsson, Paul Oakes, Jared
Richardson, Linda Rising, Johanna Rothman, Jeremy Sydik, Steph
Thompson, Derek Wade, and everyone else who posted their
thoughts, experiences, and readings. Their combined experiences
are invaluable.
Special thanks to June Kim for his many contributions throughout
the book, including pointers to far-flung research and stories of his
own experiences, as well as his feedback throughout the stages of
birthing this book.
Special thanks also to Dr. Patricia Benner, who introduced the
Dreyfus model of skills acquisition to the nursing profession, for
her support and permission to quote from her works and for her
enthusiasm for learning.
Thanks go to Dr. Betty Edwards, who pioneered the practical appli-
cations of lateral specialization in the brain, for her kind support
and permission to quote from her works.
Report erratum
Prepared exclusively for Jose Luis Loya
gggggggggggggggg
this copy is (P2.0 printing, January 2009)
GRATEFUL ACKNOWLEDGMENTS
24
Thanks to Sara Lynn Eastler for the index, to Kim Wimpsett for
correcting my many typos and often haphazard grammar, and to
Steve Peter for implementing a plethora of typesetting tricks.
And finally, thanks to you for purchasing this book and beginning
the journey with me.
Let’s move our profession forward in the right direction, harness
our experience and intuition, and create new environments where
learning matters.
Report erratum
Prepared exclusively for Jose Luis Loya
gggggggggggggggg
this copy is (P2.0 printing, January 2009)
We can’t solve problems by using the same kind of
thinking we used when we created them.
Albert Einstein
Chapter 2
Journey from Novice to Expert
Wouldn’t you like to be the expert? To intuitively know the right
answer? This is the first step of our journey together along that
road. In this chapter, we’ll look at what it means to be a novice
and what it means to be an expert—and all the stages in between.
Here’s where our story begins.
Once upon a time, two researchers (brothers) wanted to advance
the state of the art in artificial intelligence. They wanted to write
software that would learn and attain skills in the same manner
that humans learn and gain skill (or prove that it couldn’t be done).
To do that, they first had to study how humans learn.
They developed the Dreyfus model of skill acquisition,1 which out-
lines five discrete stages through which one must pass on the jour-
ney from novice to expert. We’ll take a look at this concept in depth;
as it turns out, we’re not the first ones to use it effectively.
Back in the early 1980s, the nursing profession in the United
States used the lessons of the Dreyfus model to correct their
approach and help advance their profession. At the time, the prob-
lems faced by nurses mirrored many of the same problems pro-
grammers and engineers face today. Their profession has made
great progress, and in the meantime we still have some work to
do with ours.
1.
Described in Mind Over Machine: The Power of Human Intuition and Expertise in the Era of the Computer [DD86].
Prepared exclusively for Jose Luis Loya
gggggggggggggggg
CHAPTER 2. JOURNEY FROM NOVICE TO EXPER T
26
Event Theories vs. Construct Theories
The Dreyfus model is what’s called a construct theory.
There are two types of theories: event theories and con-
struct theories.∗ Both are used to explain some phe-
nomenon that you’ve observed.
Event theories can be measured; these types of theories
can be verified and proven. You can judge the accuracy
of an event theory.
Construct theories are intangible abstractions; it makes no
sense to speak of “proving them.” Instead, construct the-
ories are evaluated in terms of their usefulness. You can’t
judge a construct theory to be accurate or not. That’s mix-
ing apples and existentialism. An apple is a thing; existen-
tialism is an abstraction.
For instance, I can prove all sorts of things about your
brain using simple electricity or complex medical imaging
devices. But I can’t even prove you have a mind. Mind is
an abstraction; there’s really no such thing. It’s just an idea,
a concept. But it’s a very useful one.
The Dreyfus model is a construct theory. It’s an abstraction,
and as we’ll see, it’s a very useful one.
∗.
See Tools of Critical Thinking: Metathoughts for Psychology [Lev97].
Here are some observations that ring true for both nurses and pro-
grammers, and probably other professions as well:
• Expert staff members working in the trenches aren’t always
recognized as experts or paid accordingly.
• Not all expert staff want to end up as managers.
• There’s a huge variance in staff members’ abilities.
• There’s a huge variance in managers’ abilities.
• Any given team likely has members at widely different skill
levels and can’t be treated as a homogeneous set of replace-
able resources.
There’s more to skill levels than just being better, smarter, or faster.
The Dreyfus model describes how and why our abilities, attitudes,
capabilities, and perspectives change according to skill level.
Report erratum
Prepared exclusively for Jose Luis Loya
gggggggggggggggg
this copy is (P2.0 printing, January 2009)
NOVICES VS. EXPER TS
27
Figure 2.1: A Unix wizard
It helps explain why many of the past approaches to software devel-
opment improvement have failed. It suggests a course of action
that we can pursue in order to meaningfully improve the software
development profession—both as individual practitioners and for
the industry as a whole.
Let’s take a look.
2.1 Novices vs. Experts
What do you call an exper
t software developer? A wizard. We work
with magic numbers, things in hex, zombie processes, and mystical
incantations such as tar -xzvf plugh.tgz and sudo gem install --include-
dependencies rails.
We can even change our identity to become someone else or trans-
form into the root user—the epitome of supreme power in the Unix
world. Wizards make it look effortless. A dash of eye of newt, a lit-
tle bat-wing dust, some incantations, and poof! The job is done.
Report erratum
Prepared exclusively for Jose Luis Loya
gggggggggggggggg
this copy is (P2.0 printing, January 2009)
NOVICES VS. EXPER TS
28
Making It Look Easy
I once was in a position to interview professional organists.
For an audition piece, I chose Charles-Marie Widor’s “Toc-
cata” (from Symphony No. 5 in F Minor, Op. 42 No. 1, for
those who care about such things), a frenetic piece that
sounded suitably difficult to my amateur ears.
One candidate really worked it—both feet flying on the
pedals, hands running up and down both ranks of the
organ in a blur, a stern look of intense concentration across
her brow. She was practically sweating. It was a terrific per-
formance, and I was suitably impressed.
But then came along the true expert. She played this dif-
ficult piece a little bit better, a little bit faster, but was smil-
ing and talking to us while her hands and feet flew in an
octopus-like blur.
She made it look easy, and she got the job.
Despite the mythological overtones, this vision is fairly common
when considering an expert in any particular field (ours is just
arcane enough to make it a really compelling image).
Consider the expert chef, for instance. Awash in a haze of flour,
spices, and a growing pile of soiled pans left for an apprentice to
clean, the expert chef may have trouble articulating just how this
dish is made. “Well, you take a bit of this and a dash of that—not
too much—and cook until done.”
Chef Claude is not being deliberately obtuse; he knows what “cook
until done” means. He knows the subtle difference between just
enough and “too much” depending on the humidity, where the
meat was purchased, and how fresh the vegetables are.
It’s often difficult for experts to explain
It’s hard to articulate
their actions to a fine level of detail; many
expertise.
of their responses are so well practiced
that they become preconscious actions.
Their vast experience is mined by nonverbal, preconscious areas
of the brain, which makes it hard for us to observe and hard for
them to articulate.
Report erratum
Prepared exclusively for Jose Luis Loya
gggggggggggggggg
this copy is (P2.0 printing, January 2009)
THE FIVE DREYFUS MODEL STAGES
29
When experts do their thing, it appears almost magical to the rest
of us—strange incantations, insight that seems to appear out of
nowhere, and a seemingly uncanny ability to know the right answer
when the rest of us aren’t even all that sure about the question.
It’s not magic, of course, but the way that experts perceive the
world, how they problem solve, the mental models they use, and so
on, are all markedly different from nonexperts.
A novice cook, on the other hand, coming home after a long day
at the office is probably not even interested in the subtle nuances
of humidity and parsnips. The novice wants to know exactly how
much saffron to put in the recipe (not just because saffron is ridicu-
lously expensive).
The novice wants to know exactly how long to set the timer on
the oven given the weight of the meat, and so on. It’s not that the
novice is being pedantic or stupid; it’s just that novices need clear,
context-free rules by which they can operate, just as the expert
would be rendered ineffective if he were constrained to operate
under those same rules.
Novices and experts are fundamentally different. They see the
world in different ways, and they react in different ways. Let’s look
at the details.
2.2 The Five Dreyfus Model Stages
In the 1970s, the brothers Dreyfus (Hubert and Stuart) began doing
their seminal research on how people attain and master skills.
The Dreyfus brothers looked at highly
skilled practitioners, including commer- Dreyfus is applicable
cial airline pilots and world-renowned per skil .
chess masters.2 Their research showed
that quite a bit changes as you move from novice to expert. You
don’t just “know more” or gain skill. Instead, you experience fun-
damental differences in how you perceive the world, how you
approach problem solving, and the mental models you form and
use. How you go about acquiring new skills changes. External fac-
tors that help your performance—or hinder it—change as well.
2.
Cited in From Novice to Expert: Excellence and Power in Clinical Nursing Practice [Ben01].
Report erratum
Prepared exclusively for Jose Luis Loya
gggggggggggggggg
this copy is (P2.0 printing, January 2009)
THE FIVE DREYFUS MODEL STAGES
30
Unlike other models or assessments that rate the whole person,
the Dreyfus model is applicable per skill. In other words, it’s a sit-
uational model and not a trait or talent model.
You are neither “expert” nor “novice” at all things; rather, you are at
one of these stages in some particular skill domain. You might be
a novice cook but an expert sky diver, or vice versa. Most nondis-
abled adults are experts at walking—we do so without planning
or thinking. It has become instinct. Most of us are novices at tax
preparation. We can get through it given a sufficient number of
clear rules to follow, but we really don’t know what’s going on (and
wonder why on Earth those rules are so arcane).
The following are the five stages on the journey from novice to
expert.
Stage 1: Novices
Novices, by definition, have little or no previous expe-
rience in this skill area. By “experience,” I mean
specifically that performing this skill results in a
change of thinking. As a counterexample, consider the case of the
developer who claims ten years of experience, but in reality it was
one year of experience repeated nine times. That doesn’t count as
experience.
Novices are very concerned about their ability to succeed; with lit-
tle experience to guide them, they really don’t know whether their
actions will all turn out OK. Novices don’t particularly want to
learn; they just want to accomplish an immediate goal. They do
not know how to respond to mistakes and so are fairly vulnerable
to confusion when things go awry.
They can, however, be somewhat effective if they are given context-
free rules to follow, that is, rules of the form “Whenever X hap
pens,
do Y.” In other words, they need a recipe.
This is why call centers work. You can hire
Novices need recipes.
a large number of folks who don’t have a
lot of experience in the subject matter at
hand and let them navigate a decision tree.
Report erratum
Prepared exclusively for Jose Luis Loya
gggggggggggggggg
this copy is (P2.0 printing, January 2009)
THE FIVE DREYFUS MODEL STAGES
31
Figure 2.2: Recipe for corn muffins. But how long do you cook it?
A giant computer hardware company might use a script like this:
1. Ask the user whether the computer is plugged in.
2. If yes, ask whether the computer is powered on.
3. If no, ask them to plug it in and wait.
4. and so on...
It’s tedious, but fixed rules such as these can give novices some
measure of capability. Of course, novices face the problem of not
knowing which rules are most relevant in a given situation. And
when something unexpected comes up, they will be completely
flummoxed.
Report erratum
Prepared exclusively for Jose Luis Loya
gggggggggggggggg
this copy is (P2.0 printing, January 2009)
THE FIVE DREYFUS MODEL STAGES
32
As with most people, I am a novice when it comes to doing my
taxes. I have little experience; despite having filed taxes for more
than twenty-five years, I haven’t learned anything or changed my
thinking about it. I don’t want to learn; I just want to accomplish
the goal—to get them filed this year. I don’t know how to respond
to mistakes; when the IRS sends me a terse and rather arrogant
form letter, I usually have no idea what they’re on about or what to
do to fix it.3
There is a solution, of course. A context-free rule to the rescue!
Perhaps it’s something such as the following:
• Enter the amount of money you earned last year.
• Send it in to the government.
That’s simple and unambiguous.
The problem with recipes—with context-free rules—is that you can
never specify everything fully. For instance, in the corn muffin
recipe, it says to cook for “about 20 minutes.” When do I cook
longer? Or shorter? How do I know when it’s done? You can set
up more rules to explain, and then more rules to explain those,
but there’s a practical limit to how much you can effectively specify
Pragmatic Thinking and Learning Page 3