ERIC EJ1137093: A Design Quality Learning Unit in OO Modeling Bridging the Engineer and the Artist

2.2 MB·English

by  ERIC

Information Systems Education Journal (ISEDJ) 13 (1)
ISSN: 1545-679X January 2015
A Design Quality Learning Unit in OO Modeling
Bridging the Engineer and the Artist
Leslie J. Waguespack
[email protected]
Computer Information Systems Department
Bentley University
Waltham, Massachusetts 02452, USA
Abstract
Recent IS curriculum guidelines compress software development pedagogy into smaller and smaller
pockets of course syllabi. Where undergraduate IS students once may have practiced modeling in
analysis, design, and implementation across six or more courses in a curriculum using a variety of
languages and tools they commonly now experience modeling in four or fewer courses in at most a
couple of paradigms. And in most of these courses their modeling decisions focus on acceptable
syntax rather than principles representing and communicating concepts of quality in information
systems. Where learning design quality may once have been an osmotic side effect of development
practice it must now be a conscious goal in pedagogy if it is to be taught at all. This paper presents a
learning unit that teaches design quality in object-oriented models. The focus on object-oriented
models allows the learning to permeate analysis, design, and implementation enriching pedagogy
across the systems development life cycle. The quality perspective presented is more expansive than
that usually found in software engineering, the traditional “objective” notion of metrics, and integrates
aspects of aesthetics, the more subjective phenomena of satisfaction. This learning unit is intended as
an adaptable framework to be tailored to the coursework and the overall objectives of specific IS
programs.
Keywords: design quality, design, OO modeling, IS discipline, IS curricula, IS pedagogy
1. INTRODUCTION constrain the scope of coursework by
compressing systems development into smaller
Over the past decade computing curricula have
and smaller pockets of course syllabi. (AACSB
been repartitioned with the permeation of
2010, EQUIS 2010) Where undergraduate IS
computing across disciplines and society.
students once may have practiced modeling in
(Shackelford, Cross, Davies, Impagliazzo,
analysis, design, and implementation across six
Kamali, LeBlanc, Lunt, McGettrick, Sloan & Topi,
or more courses in a program using a variety of
2005) There are now 5 major computing
languages and tools they commonly now
curriculum guidelines that subdivide computing.
experience modeling in four or fewer courses in
(Soldan, Hughes, Impagliazzo, McGettrick,
at most a couple of paradigms. (Waguespack
Nelson, Srimani & Theys 2004, Cassel,
2011) And in most of these courses their
Clements, Davies, Guzdial, McCauley,
modeling decisions focus on acceptable syntax
McGettrick, Sloan, Snyder, Tymann & Weide,
rather than principles representing and
2008, Diaz-Herrara & Hilburn, 2004, Lunt,
communicating concepts of quality in
Ekstrom, Gorka, Hislop, Kamali, Lawson,
information systems. Where learning design
LeBlanc, Miller & Reichgelt, 2008, Topi, Valacich,
quality may once have been an osmotic side
Wright, Kaiser, Nunamaker, Sipior & de Vreede,
effect of development practice it must now be a
2010) The co-location of IS curricula in schools
conscious goal in pedagogy if it is to be taught
of business further exacerbates the pressure on
at all.
pedagogy as accreditation bodies further
©2015 EDSIG (Education Special Interest Group of the AITP) Page 58
www.aitp-edsig.org /www.isedj.org
Information Systems Education Journal (ISEDJ) 13 (1)
ISSN: 1545-679X January 2015
At the same time industry and academia persist (Waguespack 2010b) asserts that the quality of
in their lament over the paucity of focus on systems revolves around two primary concepts:
quality in system design first sounded more than efficiency and effectiveness defined as follows
four decades ago (Dijkstra, 1968) and echoing (New Oxford American Dictionary):
consistently since as in (Denning, 2004, Brooks
Efficiency [noun]- the ratio of the useful work
1995, 2010, Beck, Beedle, van Bennekum,
performed […] in a process to the total energy
Cockburn, Cunningham, Fowler, Grenning,
[effort] expended
Highsmith, Hunt, Jeffries, Kern, marick, Martin,
Mellor, Schwaber, Sutherland, & Thomas 2010) Effectiveness [noun]- successful in producing a
desired or intended result
This paper presents a learning unit that teaches
design quality within the object-oriented These two concepts appear primarily
paradigm. The focus on OO models allows the quantitative and therefore objective. In and of
learning to permeate analysis, design, and themselves they may well be. Portraying
implementation enriching pedagogy across the efficiency using a convenient interpretation of
systems development life cycle. We amplify a “work” and “effort” is genuinely objective. “How
traditional “objective” notion of systems quality many” or “how much” or “how often” often
(i.e. metrics usually found in software depicts efficiency. But, when we ask “Is it
engineering) by integrating the more subjective enough?” apparent objectivity fades away.
phenomena of satisfaction, aesthetics. This
Likewise, the supposed objectivity of
learning unit is adaptable to the coursework and
“effectiveness” relies upon the tenuous phrase,
objectives of specific IS programs. The paper
“desired or intended result” defined as
presents: a brief overview of design quality,
properties to assess design choices, the object- Intend [noun]- have (a course of action) as
oriented ontology; and a discussion of how each one’s purpose or objective; plan
of the design choice properties express quality
through the use of object-oriented modeling Effectiveness (like efficiency) is a
constructs. Finally, there is a description of how correspondence between a system and its
the learning unit has been integrated in object- stakeholders’ intentions. Assessing effectiveness
modeling syllabi with a comment on its efficacy. depends on comparing “what is” to “what is
intended.” While the former may be expressed
quantitatively the latter presents challenges:
2. WHAT IS DESIGN QUALITY?
clarity of conception, mode of representation,
Quality is an elusive concept, shifting and scope of contextual orientation, and fidelity of
morphing on a supposed boundary between communication to name but a few. Indeed the
science and art: objective, engineering notion of effectiveness is complicated when we
characteristics versus subjective, aesthetic contemplate identifying and quantifying the
observer or stakeholder experience. stakeholder(s) intentions objectively.
International standards of quality reflect the
The indefiniteness or imprecision that
challenge of defining quality by offering a variety
characterizes stakeholder intention(s) is
of perspectives (as gathered here by Hoyle
generally not a concern if an observer is asked
2009):
to assess the beauty of something – an
 A degree of excellence (Oxford English assessment generally conceded to be subjective.
Dictionary) A detailed or even explicit intention is not
expected in assessing beauty – beauty is most
 Freedom from deficiencies or defects
often perceived as an experience of observation
(Juran 2009)
rather than a system analysis. Most people
 Conformity to requirements (Crosby commonly accept beauty as subjective and
1979) exempt from specific justification or explanation
– “Beauty is in the eye of the beholder.” and
 Fitness for use (Juran 2009)
“You’ll know it [beauty] when you see it.” This
 Fitness for purpose (Sales and Supply of absence of or difficulty in forming a quantitative
Goods Act 1994) justification of beauty is often the basis for
categorizing artifacts or processes as products of
 The degree to which the inherent
art rather than of engineering. And therein lies
characteristics fulfill requirements (ISO
the presumption that the aspects of design
9000:2005)
quality that we label objective and those we
 Sustained satisfaction (Deming 1993) label subjective are somehow dichotomous. They
in fact teeter between objectivity and

©2015 EDSIG (Education Special Interest Group of the AITP) Page 59
www.aitp-edsig.org /www.isedj.org
Information Systems Education Journal (ISEDJ) 13 (1)
ISSN: 1545-679X January 2015
subjectivity depending on the degree of graphical outline of the ontology is Figure 1
granularity that observers choose to employ in below.
inspecting not only the artifact but also their
own disposition toward satisfaction relative to it.
3. AN ARCHITECTURAL INTERPRETATION
OF QUALITY DESIGN
We will never be able to absolutely define design
quality because of the relativistic nature of
satisfaction in the observer experience. But, our
students must still face design choices. So, as IS
educators we must provide a framework for
them to develop and refine their individual
perceptions and understanding of systems
quality. The taxonomy of design choice
evaluation proposed in Waguespack (2008,
2010b), the 15 choice properties, is just such a
framework. (See Appendix A.) Choice properties
derive from Christopher Alexander’s writings on
design quality in physical architecture. Figure 1 – Object-Oriented Ontology
(Alexander 2002)
The ontology captures the elements of the
Choice properties address the process of object-oriented paradigm eschewing the
building, the resulting structure, and the obfuscation that usually occurs with
behavior of systems as cultural artifacts. Every programming language syntax examples. At the
design decision, choice, contributes to the same time an experienced IS teacher can readily
aggregate observer experience: either positively translate the ontological elements into a relevant
or negatively. Each choice exhibits the 15 programming dialect.
properties with varying strengths or influence
that impact the resulting observer satisfaction. 5. CRAFTING OBJECT-ORIENTED MODELING
The confluence of property strength results from CHOICES THAT STRENGTHEN PROPERTIES
the coincidence of the designer’s choice with the OF DESIGN QUALITY
collective intention of the stakeholders. The
combination of all choices with their respective This section, the heart of the learning unit,
property strengths results in the overall, enumerates the 15 choice properties as defined
perceived design quality. Many of the properties in Waguespack (2010b) illustrating how
are design characteristics long recognized in modeling choices in the object-oriented ontology
software engineering (i.e. modularization, can express design quality. In this space-limited
encapsulation, cohesion, etc.). But several reach discussion one choice property often references
beyond engineering to explain aesthetics, the art another reflecting the confluent nature of the
(i.e. correctness, transparency, user friendliness, design quality properties as Alexander defines
elegance, etc.). An example of the effectiveness them in physical architecture. (Alexander 2002)
of choice properties in explaining the design
Stepwise Refinement (as the name implies) is
quality of production systems is reported in
an approach to elaboration that presumes a
(Waguespack, Schiano & Yates 2010a).
problem should be addressed in stages. The
stages may represent degrees of detail or an
4. THE ONTOLOGY OF THE OBJECT- expanding problem scope. (Birrell and Ould
ORIENTED PARADIGM 1988) In either case quality evidence of stepwise
refinement is demonstrated by the cogent and
Illustrating design decisions in the object-
complete representation of a design element at
oriented paradigm can be a challenge. The
whatever level of detail or scope is set at each
idiosyncrasies of OO programming syntax often
stage. To achieve this representation the
obscure the intention and/or the result of a
modeling paradigm must support abstraction
design decision. For that reason the learning unit
that allows generalization of the scope of
presented here uses a paradigm description
interest and then the elaboration of that scope
independent of programming language, the
from one stage to the next.
object-oriented ontology, found in (Waguespack
2009) and excerpted in Appendix B. The The class concept in OO provides this capability.
Through the inheritance relationship a class can

©2015 EDSIG (Education Special Interest Group of the AITP) Page 60
www.aitp-edsig.org /www.isedj.org
Information Systems Education Journal (ISEDJ) 13 (1)
ISSN: 1545-679X January 2015
represent the more abstract, general character promote a texture exhibiting a sense of system
of a model feature while expressing all the connectedness, elegance.
information and behavior needed at that level of
Encapsulation is a design quality reflected
abstraction: 1) what responsibilities the objects
directly in the nature of the object-oriented
of this class fulfill, 2) what information they
ontology as objects encapsulate both their data
manage, and 3) what services this class’s
and behavioral attributes. Encapsulation clearly
objects provide the rest of the model. As the
delineates who is allowed to manipulate system
modeling stages progress greater specialization
information and who is not. Object data and
is achieved with child classes that redefine
behavior are only accessible (invoke-able) via
abstract behaviors: by adding data and/or
the published services defined for each object by
behavioral attributes germane only at a lower
its class. When sustained as a discipline this
level of abstraction, or by defining collaborations
boundary universally designates the object as
to support this class’s responsibilities. Stepwise
the finest granule of modularization. (Scott
Refinement can mimic the concept of “need-to-
2006) This principle eliminates the possibility of
know.” Only that detail required to “understand”
“side effects” where system state changes occur
the system at that abstraction level need be
in any manner other than the “contractual”
revealed or perhaps is not even chosen until the
prescription defined in the object’s service
need arises. When the need does arise the detail
interface. The isolation of the inside of the object
may be added within the genealogy of the class
from the outside allows both to evolve without
preserving the cohesion of a class’s defined
servitude to the implementation of the other
functional responsibility at the higher abstraction
(e.g. pursuing efficiency) as an object is
levels.
obligated only through the published
As an example, consider a class that defines responsibilities in its class’s services.
items stored in an inventory. At the most
Extensibility is the property of design quality
general level the most important functional
most important in pursuing systems with
detail is the entry and removal of items. As
sustainability essential to cost of ownership
refinement progresses simple entry and removal
economy. This is the vehicle for seamless
may be augmented by including item re-order
unfolding in system evolution. Extensibility
and supplier interaction both concealed from the
juxtaposes the potential for new functionality
inventory item’s client who sees only entry and
with the effort required to achieve it. (van Vliet
removal. The supplier interaction details are
2008). In the object-oriented paradigm class
encapsulated within the inventory item’s
plays the pivotal role by empowering instance
responsibilities retaining the cohesion of the
and inheritance relationships.
class’s purpose (its identity). And the description
of the inventory item exhibits correctness at Multiplicity is achieved through instance
either level of detail with and without the propagation, progeny. Each instance is
supplier interaction elaboration. completely interoperable in any combination
with its sibling objects as well as acting as an
Cohesion is a quality property reflecting a
instance of any ancestor class.
consistent responsibility distribution in a field of
Interchangeability both enables and reinforces
system components. (Zuse 1997) Since every
modularization.
object “expects” the objects around it to fulfill
their responsibilities to contribute to the whole Evolution or unfolding is accomplished as class
model, each object is in itself free to be single- definitions are refined and specialized in their
minded in its focus on its own purpose. This is child classes – the relationship called
the result of well-chosen classes. This inheritance. When a child class extends the
independent sufficiency accentuates the scope of the data and behavioral attributes of its
divisibility of function in terms of each object’s parent it honors the pattern set out in the parent
individual purpose, its identity, and the clarity without contradiction. Polymorphism
with which its purpose is exposed to the rest of compensates (through dynamic binding) for any
the community of objects in the system. The overridden methods. This extension proceeds
single-mindedness that results also increases without any impairment of correctness because
the feasibility of object interaction the interfaces defined in the parent class must
rearrangement enabling an overall change in be supported in each child class. The parent to
system function while almost every class’s child unfolding specializing structure and
individual purpose remains fixed. The behavior results in an unbroken thread that
independent sufficiency of each object’s inner binds each class to its ancestry and projects an
workings couples with the system-wide identity down through the generations of class.
interdependency of object cooperation to

©2015 EDSIG (Education Special Interest Group of the AITP) Page 61
www.aitp-edsig.org /www.isedj.org
Information Systems Education Journal (ISEDJ) 13 (1)
ISSN: 1545-679X January 2015
Modularization along with cohesion expresses Transparency is evident structure, revealing
“divide and conquer” problem solving how things fit and work together. (Kaisler 2005)
augmented by the flexibility of configuring and In the OO ontology “fit together” and “work
reconfiguring objects as cooperating agents. together” are defined by the structural and
Modularization also supports scale permitting the behavioral relationships. Individual objects may
composition of subsystems of varying scope that represent clearly delineated and encapsulated
hold details in abeyance until they require focus. choices, but their cooperation is defined by
(Baldwin and Clark 2000) Enlightened module relationships.
design exposes the solution structure envisioned
Inheritance explains the structural relationship
by the modeler and publishes intentions for
of classes through the propagation of data and
further extension by separation of concerns and
behavioral attributes. Inheritance not only
isolation of accidents of implementation. (Brooks
propagates attributes, but also enables a class
1987) The OO paradigm provides ample facility
hierarchy’s capacity for exhibiting similarity and
for defining modules of any size and scope while
difference between parent and child classes.
aggregating and/or nesting their interfaces
That which is similar (in fact identical) inherited
through deliberate information hiding. The
by the child class is assumed and becomes in
granularity enabled through modularization may
effect familiar – requiring no reiteration. This
be applied to facilitate the modeler’s formulation
“folding” of that which is not changed avoids
of structure as well as the perspective to aid
clutter in the child class description, but may be
stakeholder recognition and understanding.
readily reviewed in the parent.
Correctness in software engineering is often
The behavioral relationships of association,
narrowly defined as computing the desired
message passing, and polymorphism explain the
function. (Pollack 1982) Thriving Systems
predictable patterns of communication and
Theory frames this property upon two outcomes:
action. Association uses the property of identity
1) validation, the clarity and fidelity of the
to designate membership, ownership, and
represented understanding of system
accessibility among objects. Message passing
characteristics, and 2) verification, the
provides the mechanism for cooperating action
completeness and effectiveness of model feature
between objects providing a disciplined conduit
testing both individually and in composition.
through the encapsulating boundary of objects
Validation depends on the fidelity of the by using services to convey intention,
unfolding process; that through the stages of information, and reaction. Polymorphism allows
stepwise refinement the “essence” of system the abstraction of intention by using the same
characteristics are brought forward maintaining service name to evoke distinct behaviors from
their integrity. (Brooks 1987) Modularization objects of different classes. The identical service
aids in cataloging and focusing on individual names in classes with different methods directly
essential characteristics. Correctness is the only realize the metaphorical abstraction of object
choice property that directly supports itself! behavior where at one level of abstraction the
Correctness must be a priority at each stage as behaviors are the same and at a more detailed
experience shows that correctness shortcomings level of abstraction their behaviors are distinct.
grow more and more expensive to rehabilitate
Composition of Function - As a fundamental
as evolution progresses – notice “rehabilitate,”
tool for managing complexity humans regularly
to restore to normal life.
attempt to decompose problems, issues, or
Verification depends on the effective testability tasks into parts that either in themselves are
of each choice to certify it as “consistent with sufficiently simple to permit direct solution or
stakeholder understanding.” Modularization can through recursion be subdivided
enables the verification of individual choices or successively until they become sufficiently
modules. Then relying on the correctness inside simple. This is a defining aspect of
modules verification can turn to the certification modularization. When the conception of a part
of behaviors resulting from composition of also anticipates reuse then the part takes on a
function. Experience often leads to dependable larger significance. The combination of
patterns of classes or modules applicable or specifying a choice consistent with the essence
adaptable to recurring modeling tasks. of system characteristics and then designing the
Verification in these situations can focus on choice as an interchangeable component in
known areas of fragility/risk limiting the effort multiple super-ordinate choices is a step toward
required to reach a desired confidence level of elegance. Reusable choices represent an
reliability. understanding of the essence of the system at a
deeper level than an individual application. They

©2015 EDSIG (Education Special Interest Group of the AITP) Page 62
www.aitp-edsig.org /www.isedj.org
Information Systems Education Journal (ISEDJ) 13 (1)
ISSN: 1545-679X January 2015
represent awareness of the intention, perhaps (association and message passing) each object
even the philosophy of the system domain. must be uniquely distinguishable.
Composition of function as a property of design Scale’s effect on design quality is reflected in
quality is realized in model features that common idioms: “You can’t see the forest for
facilitate the extension or retargeting of the the trees!” and “Let’s get a view from 10,000
model in the future. It is the capacity to combine feet.” They reflect the importance of context in
simple functions to build more complicated ones recognition and decision-making. Scale captures
(Meyer 1988). The retargeting capability may be the modeling imperative that all choices must be
provided directly to the users of the system in kept in perspective because it is not sufficient to
the form of a programmable interface. A choice consider a choice only in the microcosm of itself,
achieving the principle of composition of function as it must also participate in the connectedness
is marked not only by the function it initially of the whole. By achieving scale, a system
provides the user, but also by the functionality it designer provides differing granularities of
anticipates and supports even (perhaps) before comprehensibility to suit the requirements of a
the stakeholders realize the need for the variety of observers (Waguespack 2010).
capability.
The relationships provided in the object-oriented
Identity is at the root of recognition and is paradigm (association, inheritance, instance,
another property of design quality not usually message passing, and even polymorphism)
defined in software engineering. In the physical provide ample means for designing collections of
world identity is literal based upon direct cooperating choices that are nested, intersect,
sensorimotor experience: by sight or touch and or partition the full field of functionality essential
in some cases by sound or smell – a human to the model. These may be called variously
experience of the “real” world. In the object- subsystems, modules, or sub-modules. In those
oriented paradigm identity is an object property. cases where the actual structure of a collection
(Khoshafian and Copeland 1986) Existence is must be rendered obscure, classes and objects
sufficient for object identification. can be devised to serve as facades or agents to
“keep up appearances.” Coupled with stepwise
In other paradigms identification is achieved
refinement, as it is, scale is used to focus
through possessed characteristics (attributes)
modeler and stakeholder attention to achieve
that contribute to distinct recognition by a
the contextual understanding needed to address
process of intersecting categorizations or the
constituent concerns within the whole.
introduction of an artificial characteristic whose
sole purpose is to support discrimination. Aside User Friendliness is another property of design
from the fact that these approaches to quality more often considered aesthetic. It is a
identification require some overhead (either combination of: ease of learning; high speed of
mental or computational) they are simply not user task performance; low user error rate;
natural to humans. Humans perceive objects as subjective user satisfaction; and, user retention
possessing characteristics rather than over time (Shneiderman 1992). Its impact may
characteristics defining objects. The former be easiest to consider in its absence. A modeling
begins with certain uniqueness and progresses choice that is “unfriendly” to stakeholders is
toward explanation while the latter begins with confusing, hard to comprehend, unwieldy, and
uncertainty and attempts to deduce uniqueness. perhaps worst of all, of indeterminate
correctness. That which defies understanding
Characteristics are not unimportant.
cannot be determined to be correct. Satisfaction
Classification is essential in most human
is cumulative. The sensitivity to the
problem solving activities. And recognition is
stakeholders’ conceptions of the essence of the
virtually always accelerated by the
system to be modeled is key to the stakeholders’
discrimination that categorizing characteristics
sense of comfort, familiarity, and expectation.
(attributes) provide. And most importantly in the
absence of physical experience categorization The object-oriented paradigm excels in its
through characteristics is the only choice. Class facility to represent systemst preserves the
structure and the instance relationship are vital stakeholders’ ability to recognize “their” system.
to identity – an object belongs to “this” class Authoring object-oriented models whose
and not to “another.” Described both by what an elements correspond almost one-to-one with the
object “knows” (data attributes) and what it real-world concepts and entities results in
“knows how to do” (behavioral attributes) intrinsically better stakeholder understanding
classes form a categorization cornerstone of the and interaction. The casting of “objects” in the
object-oriented ontology. But to model both the models that have direct counterparts in the
static and dynamic dimensions of reality stakeholders’ experience exhibits a

©2015 EDSIG (Education Special Interest Group of the AITP) Page 63
www.aitp-edsig.org /www.isedj.org
Information Systems Education Journal (ISEDJ) 13 (1)
ISSN: 1545-679X January 2015
fundamentally friendly quality. It respects the Programmability in software engineering is
stakeholders’ perceptions and it welcomes them often considered a feature rather than a
into the processes of verification and validation property of design quality – the capability within
that are intrinsic to correctness. The unified hardware and software to change; to accept a
structure of “what an object knows” and “what new set of instructions that alter its behavior
an object knows how to do” correlates so (Birrell and Ould 1988). It is closely allied with
naturally with observers of business models or extensibilityand addresses the need for models
process models that the natural clarity in that to welcome the future. What largely separates
communication improves understanding and information systems from other human-made
avoids mistakes in understanding, mechanisms is the degree of adaptability that
communication, or implementation. they offer to deal gracefully with change. Unlike
most appliances that support a very narrow
And in a serendipitous quirk of language (or a
range of use (albeit with great reliability),
profound emergence of the deep meaning of
contemporary information systems are expected
metaphors) Alexander’s term from which the
to provide not only amplification of effort as in
principle here, user friendliness, is derived is
computation, but also amplification of
roughness. (Alexander 2002) Something has to
opportunity in terms of different approaches to
have a certain degree of roughness if one is to
business or organizational questions.
be able to effectively grasp it!
Contemporary information systems are expected
Patterns describe versatile templates to solve to demonstrate that they can reliably
particular problems in many different situations accommodate change. As with extensibility,
(Gamma et al. 1995). All entities in the object- successful accommodation of change relies on
oriented paradigm propagate from classes, an understanding of the fundamental options
predefined templates, or “cookie cutters.” This governing the structure and behavior within a
protocol organizes what otherwise would be a particular domain. The OO ontology offers
bewildering multiplicity of individual powerful tools (structural and behavioral
computational entities to consider. It becomes relationships, e.g. inheritance and
less complicated in the understanding that the polymorphism) to service the elements of
potential of any number of objects boils down to change without fracturing a skeletal foundation
understanding the class(s) of which they are of base classes characterizing the domain.
instances. Each instance mimics perfectly the
What sets programmability apart from
form and function of every other of its siblings,
extensibility is a facility that permits altering the
members of that class. Class hierarchies,
systems behavior without having to reconstruct
generations of parent-child class definitions,
choices – that is to say that the system’s
defining “nearly the same” and “different in
behavior can be sensitive to the context
specific ways” relationships significantly lessen
determined by a “user” in “real time.” “Real
the apparent complexity that considering only
time” is relative to the “user’s” role (e.g.
individual entities entails. Class hierarchies
developer or end user, etc.). This versatility is
define the path of unfolding for all to see – a
not accidental but architectural. Choices may
depiction of the analysis, solution, and design
provide an interface language for end users that
philosophies at work.
permits selections of system actions to meet an
Patterns is the property of design quality that immediate “real-time” need – an interface as
channels change (unfolding). A pattern simple as a light switch or as complex as a
foreshadows where and how change will need to natural language.
be accounted for. Patterns of the form
Reliability is a property of design quality more
popularized in (Coplein, 1995) document
often associated with implementation than
commonly encountered design questions offering
design. It is the assurance that a product will
carefully considered advice and cautions. Their
perform its intended function for the required
patterns are paradigm and modeling language
duration within a given environment (Pham
independent. However, it is not surprising that
2000). Objects facilitate modularized testing and
many examples using patterns are presented in
quality assurance. A certified class produces
OO dialects. The reason is simple. The
certified objects (which is not to say that
integration of instance, inheritance, message
certification is easy or inexpensive). As long as
passing, and polymorphism relationships is an
classes are protected from dynamic modification
ideal toolset for expressing patterns with a
in deployment there is no need to be concerned
balance of prescription and adaptability – a
with the inner workings of their objects. As long
balance not as conveniently achieved in dialects
as objects are truly encapsulated they conform
based on pre-object-oriented paradigms.
to the intention of their class. In development

©2015 EDSIG (Education Special Interest Group of the AITP) Page 64
www.aitp-edsig.org /www.isedj.org
Information Systems Education Journal (ISEDJ) 13 (1)
ISSN: 1545-679X January 2015
testing proceeds incrementally as new classes field effect of the beneficial, integrated, mutual
are added or rearranged in their collaboration. support of strong choices described in Thriving
Once deployed testing is relegated to their Systems Theory. (Waguespack 2010b)
interactions rather than their definition. Testing
is compartmentalized and does not explode 6. INTEGRATING THE DESIGN QUALITY
exponentially when additional classes or LEARNING UNIT IN AN OBJECT-ORIENTED
functionality within a class is added. MODELING SYLLABUS
Reliability in design reflects an austerity that For the past six semesters the design quality
confines design elements to the essentials of the learning unit presented here is woven into two
stakeholder’s intentions. When design or object-oriented modeling syllabi: 1)
implementation decisions involve additional undergraduate systems analysis and design and
constructs due to technology or compatibility, 2) masters level object-oriented systems
these accidents of implementation must be engineering. The unit content appears
clearly delineated so as not to imply that they throughout the pedagogy of modeling using
are essence rather than accident. This clear UML-2 syntax.
distinction will protect future system evolution
After initially presenting the object-oriented
from mistaking accidental “baggage” as
paradigm using the ontology to establish its
stakeholder intentions.
vocabulary (see Appendix B), we present use
Elegance is perhaps the epitome of subjective case, class, and sequence diagramming
quality assessment that clearly sets choice establishing the syntax and the expression of
properties of design quality apart from semantics in UML-2. During this UML
traditional software engineering metrics. presentation we repeatedly allude to the design
“Pleasing grace and style in appearance or quality properties through the syntax. Small
manner,” that’s how the dictionary expresses student groups and then individuals conduct a
the meaning of “elegance”. (Oxford English series of modeling exercises based on
Dictionary) requirement narratives establishing the
students’ grasp of UML syntax. On that
“A designer knows he has achieved
foundation the explanation of design quality, the
perfection not when there is nothing left to
enumeration of the fifteen properties, and the
add, but when there is nothing left to take
corresponding application of OO ontology
away.” (Raymond 1996)
elements to strengthen the properties precede a
Models composed of choices that are consistent, final individual course modeling project. The
clear, concise, coherent, cogent, and design quality discussion provides a quality
transparently correct exude elegance and vocabulary for one-on-one consultations
nurture cooperation, constructive criticism and between teacher and student as each develops
stakeholder community confidence. These are the object-model of their final project. In this
models that confess to their own shortcomings one-on-one context each student’s specific
because their clarity obscures nothing, even design decisions are discussed and evaluated in
omissions. These are models that satisfy relationship to the design quality properties, an
stakeholders. They appear “intuitively obvious.” opportunity for individualized, reinforced
The clarity of their composite structure is so learning and/or suggested improvements.
self-evident that they seem “simple.” The use of
The deeper subtleties of design quality present a
the OO paradigm to construct a collection of
challenge for some students particularly in a
“building blocks” in the form of a class library to
compressed format. The “light doesn’t go on”
encapsulate architectural design decisions
right away for all students. However, the
facilitates this impression of what is “intuitively
integration of the ontology and design quality
obvious.” Using well-conceived library elements
property based vocabulary establishes a
becomes so second nature, so natural, that the
touchstone that returning students report helps
builder perceives the blocks as the natural
them “to name” the “quality elements” they
primitives of construction rather than
rediscover in succeeding coursework and
constructed artifacts.
professional practice.
Elegance largely proceeds from the efficient and
In your own curricular situation the distribution
effective representation of essential system
of learning unit elements may span more than
characteristics along with those features
one course (some addressed in OO
emerging out of design decisions, accidents of
programming, requirements engineering, or
implementation, that are laid out with equal
database design, etc.), be rearranged to suit
clarity for separate consideration. This is the

©2015 EDSIG (Education Special Interest Group of the AITP) Page 65
www.aitp-edsig.org /www.isedj.org
Information Systems Education Journal (ISEDJ) 13 (1)
ISSN: 1545-679X January 2015
your modeling tools, or be adjusted to your Coplien J and Schmidt D (Eds) (1995). Pattern
course sequencing with context-appropriate Languages of Program Design, Addison-
examples. Regardless, the learning unit Wesley, Reading, MA, USA
components are flexible and robust enough to
Crosby,P. B., (1979) Quality is Free, McGraw-
suit various specific program needs.
Hill, New York, NY, USA.
7. ACKNOWLEDGEMENTS Dijkstra, E. (1968). “GOTO Statement
Considered Harmful.” Communications of the
Thanks to helpful referees. Special thanks are
ACM, 11(3), 147-148
due my colleagues David Yates and Bill Schiano
at Bentley University for their insightful Diaz-Herrara, J.L., & Hilburn, Thomas B. (eds.)
discussions and comments on these ideas. (2004). Software Engineering 2004:
Curriculum Guidelines for Undergraduate
8. REFERENCES Degree Programs in Software Engineering,
AACSB (2010). Eligibility Procedures and IEEE Computing Society (IEEE-CS),
Accreditation Standard for Business Association of Computing Machinery (ACM)
Accreditation. Retrieved July 16, 2010 from
Deming, W. E. (1993), The New Economics for
http://www.aacsb.edu/accreditation/AAACSB
Industry, Government, Education (2ed),
-STANDARDS-2010.pdf
Cambridge Press: MIT, Cambridge, MA, USA
Alexander C, (2002). The Nature of Order An
Denning, P. J. (2004). “The Great Principles of
Essay on the Art of Building and the Nature
Computing,” Ubiquity, 4(48), 4–10
of the Universe: Book I - The Phenomenon
of Life, Berkeley, California: The Center for EQUIS (2010). EQUIS Standards and Criteria.
Environmental Structure, p. 119 Retrieved July 16, 2010 from
http://www.efmd.org/attachments/tmpl_1_a
Baldwin, C. Y., and Clark, K. B. (2000). Design
rt_041027xvpa_att_080404qois.pdf
Rules, Volume 1: The Power of Modularity.
The MIT Press, Cambridge, MA. Gamma, E., Helm, R., Johnson, R., and
Vlissides, J. (1995). Design Patterns:
Beck K., Beedle M., van Bennekum A., Cockburn
Elements of Reusable Object-Oriented
A., Cunningham W., Fowler M., Grenning J.,
Software. Addison-Wesley, Reading, MA.
Highsmith J., Hunt A., Jeffries R., Kern J.,
Marick B., Martin R.C., Mellor S., Schwaber Hoyle, D. (2009). ISO 9000 Quality Systems
K., Sutherland J., & Thomas D. (2010). Handbook. Butterworth-Heinemann
Manifesto for Agile Software Development. (Elsevier); 6 ed. Burlington, MA, USA
Retrieved July 12, 2010 from
ISO 9000 (2005), http://www.iso.org/iso/qmp
agilemanifesto.org
Juran, J. M., (1999). Quality Control Handbook
Birrell, N. D., and Ould, M. A. (1988). A Practical
(6ed), McGraw-Hill, New York, NY, USA
Handbook for Software Development.
Cambridge University Press, Cambridge, UK. Kaisler, S. H. (2005). Software Paradigms.
Wiley-Interscience, Hoboken, NJ.
Brooks F. P. (1987), "No Silver Bullet: Essence
and Accidents of Software Engineering," Khoshafian, S. N., and Copeland, G. P. (1986).
Computer, Vol. 20, No. 4, pp 10-19. “Object identity,” Proceedings of ACM
Conference on Object Oriented Programming
Brooks, F. P. (1995). The Mythical Man-Month:
Systems Languages and Applications,
Essays on Software Engineering (2ed).
Portland, OR, November 1986, 406-416.
Addison-Wesley, Boston, MA.
Lunt, B.M., Ekstrom, J.J., Gorka, S., Hislop, G.,
Brooks, F. P. (2010). The Design of Design:
Kamali, R., Lawson, E., LeBlanc, R., Miller,
Essays from as Computer Scientist. Addison-
J., & Reichgelt, H. (eds.) (2008).
Wesley, Pearson Education, Inc., Boston,
Information Technology 2008: Curriculum
MA.
Guidelines for Undergraduate Degree
Cassel L., Clements A., Davies G., Guzdial M., Programs in Information Technology,
McCauley R., McGettrick A., Sloan B., Association of Computing Machinery (ACM),
Snyder L, Tymann P., & Weide B.W., IEEE Computing Society (IEEE-CS)
(2008). Computer Science Curriculum 2008
Meyer, B. (1988). Object-oriented Software
An Interim Revision of CS2001. Association
Construction. Prentice Hall, New York, NY.
of Computing Machinery (ACM), & IEEE
Computing Society (IEEE-CS)

©2015 EDSIG (Education Special Interest Group of the AITP) Page 66
www.aitp-edsig.org /www.isedj.org
Information Systems Education Journal (ISEDJ) 13 (1)
ISSN: 1545-679X January 2015
Pham, H. (2000). Software Reliability. Springer, Topi, H., Valacich, J.S., Wright, R.T., Kaiser,
Berlin, Germany. K.M., Nunamaker, J.F. Jr., Sipior, J.C., & de
Vreede, G.J. (eds.) (2010). IS2010:
Pollack, S. (Ed.). (1982). Studies in Computer
Curriculum Guidelines for Undergraduate
Science. Mathematical Association of
Degree Programs in Information Systems,
America, Washington, DC.
Association for Computing Machinery (ACM),
Raymond, E. S. (1996). The New Hacker's Association for Information Systems (AIS)
Dictionary, 3rd ed. The MIT Press,
Van Vliet, H. (2008). Software Engineering:
Cambridge, MA.
Principles and Practice, 3rd ed. Wiley,
Sales and Supply of Goods Act 1994, Ch 35, Hoboken, NJ.
Legislation of Her Majesty’s Government,
Waguespack, L. J. (2008). “Hammers, Nails,
The National Archives, UK,
Windows, Doors and Teaching Great
http://www.legislation.gov.uk/ukpga/1994/3
Design,” Information Systems Education
5/introduction
Journal, 6 (45). http://isedj.org/6/45/.
Scott, M. L. (2006). Programming Language ISSN: 1545-679X
Pragmatics, 2nd ed. Morgan Kaufmann,
Waguespack, L. J. (2009). “A Two-Page “OO
Maryland Heights, MO.
Green Card” for Students and Teachers,”
Shackelford, R., Cross, J.H., Davies, G., Information Systems Education Journal,
Impagliazzo, J., Kamali, R., LeBlanc, R., 7 (61). http://isedj.org/7/61/. ISSN: 1545-
Lunt, B., McGettrick, A., Sloan, R., & Topi, 679X
H., (2005). Computing Curricula 2005: The
Waguespack, L. J., Schiano, W. T., Yates, D. J.
Overview Report, Association for Computing
(2010a). “Translating Architectural Design
Machiner (ACM), The Association of
Quality from the Physical Domain to
Information Systems (AIS), The Computer
Information Systems,” Design Principles and
Society (IEEE-CS)
Practices: An International Journal, 4, 179-
Shneiderman, B. (1992). Designing the User 194
Interface: Strategies for Effective Human-
Waguespack, L. J. (2010b). Thriving Systems
Computer Interaction, 2nd ed. Addison-
Theory and Metaphor-Driven Modeling,
Wesley, Reading, MA.
Springer, London, U.K.
Soldan, D., Hughes, J.L.A., Impagliazzo, J.,
Waguespack, L. (2011). “Design, The “Straw”
McGettrick, A., Nelson, V.P., Srimani, K., &
Missing From the “Bricks” of IS Curricula,”
Theys, M.D. (eds.) (2004). Computer
Information Systems Education Journal, 9(2)
Engineering 2004: Curriculum Guidelines for
pp 101-108. http://isedj.org/2011-9/ ISSN:
Undergraduate Degree programs in
1545-679X
Computer Engineering, IEEE Computer
Society (IEEE-CS), Association for Zuse, H. (1997). A Framework of Software
Computing Machinery (ACM) Measurement. Walter de Gruyter, Berlin,
Germany.

©2015 EDSIG (Education Special Interest Group of the AITP) Page 67
www.aitp-edsig.org /www.isedj.org