From the Principle of Bijection to the Isomorphism of Structures: An Analysis of Some Teaching Paradigms in Discrete Mathematics


Изоморфизм структур :

Анализ некоторых аспектов преподавания

парадигмы в дискретной математике.

Кристина Эберт , Делавэр (США)

Гэри Эберт , Делавэр (США)

Михаил Клин, Беэр-Шева (Израиль)


Математика для студентов бакалавриата университета. Два -три

десятилетия назад этот курс стал требованием для математики и

информатики для  студентов-математиков  в большинстве университетов по всему миру .

Сегодня этот курс преподают  студентам  многих других дисциплин.

Работа начинается с обсуждения нескольких тем, которые

мы полагаем, должны быть включены в учебный план для любого курса в

дискретной математике и  зависит от состава аудитории.   Кроме  этого, мы

обсуждаем  несколько потенциальных моделей для преподавания курса ,

в зависимости от интересов и математического фона

аудитории .


Christine Ebert, Delaware (USA)
Gary Ebert, Delaware (USA)
Mikhail Klin, Beer Sheva (Israel)



Abstract: This paper is concerned with the teaching of Discrete
Mathematics to university undergraduate students. Two to three
decades ago this course became a requirement for math and
computer science students in most universities world wide.
Today this course is taken by students in many other disciplines
as well. The paper begins with a discussion of a few topics that
we feel should be included in the syllabus for any course in
Discrete Mathematics, independent of the audience. We then
discuss several potential models for teaching the course,
depending upon the interests and mathematical background of
the audience. We also investigate various educational links with
other components of the curriculum, consider pedagogical
issues associated with the teaching of discrete mathematics, and
discuss some logistical and psychological difficulties that must
be overcome. A special emphasis is placed on the role of
Kurzreferat: Dieser Artikel beschäftigt sich mit dem Lehren
diskreter Mathematik an einer Universität. Vor zwanzig bis
dreissig Jahren wurde solch ein Kurs Pflicht für Mathematikund Informatikstudenten. Heutzutage wird er auch von
Studenten anderer Fachrichtungen belegt. Dieser Artikel
beginnt mit der Diskussion einiger Punkte, die unserer Meinung
nach im Lehrstoff eines jeden Kurses in diskreter Mathematik
enthalten sein sollten, unabhängig von der Zuhörerschaft.
Anschliessend diskutieren wir verschiedene Modelle diesen
Kurs zu unterrichten. Diese hängen vom Interesse und dem
mathematischen Ausbildungsstand der Zuhörer ab. Wir
untersuchen ebenfalls verschiedene Verbindungen mit anderen
Komponenten des Lehrplans. Dazu gehören sowohl
pädagogische Sachverhalte in Verbindung mit diskreter
Mathematik, als auch logistische und psychische
Schwierigkeiten, die überwunden werden müssen. Spezielle
Betonung wurde auf die Rolle des Lehrbuchs gelegt.
ZDM-Classification: N75, K25
1 Introduction
Today mathematics education as a research domain is a
well-established interdisciplinary field of science. It
borders mathematics itself on one side, and philosophy
and social science on the other side (see [SieK98]). The
main target of this science is pre-college mathematics
education (kindergarten through grade 12 in the USA). At
this level scientists are involved in various kinds of
activities, ranging from concrete teaching strategies to
global politics. A recent paper [Cuo03] provides a nice
survey on the constant struggle for the improvement in
mathematics education in U.S. public schools.
However, as soon as the target shifts to the college or
university level, the investigations become much more
narrowly defined and oriented primarily on
methodological issues in certain areas of mathematics,
such as algebra, geometry, and calculus. This is
paradoxical at first sight. Indeed, the current
mathematical community consists of thousands of
professionals, most of them continually adding to their
collective experience in the teaching of university level
mathematics. While most university professors are happy
to share with one another interesting problems, new
proofs of known results, or even novel approaches to
certain topics, almost none wants their teaching
methodology questioned or reviewed. This is somehow
viewed as an infringement on “academic freedom”.
Unfortunately, relatively few seem to have much interest
in pedagogical issues in general.
This paper is concerned with the teaching of Discrete
Mathematics (DM, for short) to undergraduate students.
During the last few decades, DM has become a
recognized area within mathematics. Simultaneously, DM
has become an important part of undergraduate (and
graduate) mathematics education. In the early stages, this
process was accompanied by visible attention in the
literature. The Sloan Foundation in the U.S. funded
several colleges and universities to develop a two year
curriculum that balanced continuous and discrete
mathematics for entering students. In particular, there
were a lot of papers devoted to the teaching of DM as
part of the development of a general curriculum for
computer science students.
Our attempt here is to resume this discussion of the
educational aspects of DM. In this article we try to attract
the reader’s attention to the following general questions:
• What is the essence and spirit of DM?
• Which models for courses in DM appear most
• How does one deal with teaching DM to audiences
with widely differing interests and mathematical
• What are the “meta-goals” in teaching DM?
• What specific skills should students acquire when
taking DM?
• How does one handle the associated logistical
problems in teaching DM (grading strategy, exam
structure, textbook selection, teaching assistant roles,
Although we do not pretend to provide complete,

convincing answers to all the posed questions, we do
hope that the somewhat unusual combination of our
teaching interests and experiences allows us to provoke
some much needed dialogue on these issues.
We address this paper to two kinds of readers. On the
one hand, we hope that experts in the teaching of DM
will find the entire article ofgreat interest, including
Section 2 where we describe a few “gourmet meals” from
our “teaching kitchen”. On the other hand, for a general
mathematical educator, probably Sections 4 and 5 will be
of most interest. These sections describe the place and
role of DM in the globalpicture of mathematical
The paper consists of six sections. In Section 2 a few
topics from a standard syllabus in DM are reviewed.
Sections 3 and 4 form a global outline of various 

educational aspects of DM. In particular, we discuss
common and distinct features in various models for DM
courses, and we look at important links with other parts
of the mathematical curriculum. Section 5 is devoted to
what may be called “teaching logistics’’. Concluding
remarks are found in Section 6.
2 A Few Topics from the Syllabus
In this section we briefly outline a few important topics
from a course in DM which belong to its “kernel”; that is,
which typically appear in almost every syllabus
independent of the institution, the specific audience, or
the level of experience of the students. This kernel
includes set theory and logic, binary relations and
functions, integers, mathematical induction, elementary
combinatorics, recursions and recurrence relations, and
elements of graph theory. These topics can be taken as
our definition of DM.
2.1 Set Theory and Logic
Set theory and logic comprise the backbone of any course
in DM. In many universities worldwide, these topics,
together with relations and functions, form the basis for
an introductory course in mathematics that is obligatory
for mathematics and computer science majors as well as
for students with majors in engineering and economics.
This course may be taught over two semesters, where the
above-mentioned topics form an integral part of the first
semester. However, even if taught in one semester (with
either 3 or 4 credit hours), any serious syllabus for such a
required course will contain a non-trivial introduction to
sets and logic.
An obvious question is what should be considered first:
sets or logic? It seems to us that there is no simple answer
to this question. One can make an argument for either
order. On the one hand, beginning with logic is less
formal because the initial notion of a simple statement (or
proposition) may be introduced with the aid of several
convincing and interesting examples from both real life
and mathematics. Taking this approach, the central
logical connectives appear as a natural formalization of
the “traditional” rules of logic.
On the other hand, at least for future mathematicians, it
might be more suitable to begin with the elements of
naïve set theory as a background for all the mathematical
constructions that they might later study. Students learn
to accept the concept of a set as a fundamental idea which
cannot be formulated in terms of simpler ones, but one
which may be explained through analogies and detailed
examples. The instructor then quickly switches to a more
formal presentation, thereby providing an opportunity to
meet one of the meta-goals of DM: introducing students
to the language of mathematics. In any case, logic and set
theory should go together, in a sense “shoulder to
shoulder” (see examples below).
The main laws of set theory (more rigorously, the
axioms of Boolean algebra) must be dealt with early on.
Ideally, a teacher should introduce all the axioms
(including the absorption laws), although consideration of
most of the axioms at the very beginning is probably
sufficient. It is important to clarify as soon as possible the

“rules of the game”: namely, all other equivalences in
Boolean algebra should be proven based on these given
At this stage the instructor has an opportunity to
discuss the isomorphism between set-theoretical and
logical interpretations of the axioms of Boolean algebra.
For a more advanced audience (where a deep course in
linear algebra is a prerequisite) one may actually use this
term, appealing to the students’ experience with
isomorphisms between various models of linear spaces.
However, as a rule, one should initially avoid a strict
explanation of isomorphisms, using instead a nice
analogy between disjunction, conjunction and negation in
logic on the one hand, and intersection, union and
complementation in set theory on the other hand.
Proofs in set theory (and in logic) provide a natural
training ground for introducing students to the activity of
proving theorems in general. For example, proving the
equality of two given sets is a particularly good place to
begin. Here we distinguish four kinds of proofs.
I. Proof by Definition According to the definition, two
sets A and B are defined to be equalif and only if
each is a subset of the other. In the early stages, this
may create some difficulty for beginning students,
due to the formal nature of such proofs. However, this
is a good introduction to formal proofs in other areas
of mathematics.
II. Naïve or Intuitive ProofsThe discussion now turns to
proofs via Euler-Venn Diagrams. Advantages of this
approach are evident; namely, students have a visual
method for proving results concerning set theory and
logic. The two main disadvantages are these: first, this
approach is practically restricted to at most three
variables, although some artificial extensions may be
designed for, say, 4 variables (see [Gru84a],
[Gru84b]); second, this method depends on the
comparison of two diagramsand deciding when they
are identical (which requires some careful
formalization which is beyond the scope of DM).
III. “Algebraic” Proofs By this terminology, we mean the
use of step-by-step transformations, each one using
one of the axioms of Boolean algebra (or a previously
proven result) as a means of justification.
IV. Use of Truth Tables We refer to [Gri94], pp. 163-165, where this method is called the use of
membership tables. A justification using truth tables
can be compared with a justification using EulerVenn diagrams, establishing a natural bijection
between the cells of the diagram and the rows of the
truth table. In this way, we are getting very close to
the real notion of an isomorphism between the two
main models of Boolean algebra, although once again
the use of the term isomorphism is not needed at this
For historical accuracy it might be worthwhile to mention
the difference between Euler and Venn diagrams. Euler
introduced his famous circles in 1761 in order to visually
illustrate classical rules of logic and syllogistic reasoning.
In his diagrams with two circles you might find disjoint
circles, properly intersecting circles, or one circle
completely contained in the other, depending upon which

form of reasoning you were attempting to illustrate. Over
100 years later, Venn revised Euler’s method by using
diagrams to “prove’’ various identities in Boolean
algebra. In Venn’s approach the circles representing
independent variables were always drawn as intersecting
in the most general way. In particular, a Venn diagram
with k circles has 2k regions, and thus can be shaded in 22k 

different ways. An interesting discussion of the pros and

cons of using diagrams as a proof technique may be
found in [Gar82]. For an excellent review of all
methodological, historical, philosophical, and
mathematical issues related to the use of Euler and Venn
diagrams, we recommend the monograph [Shi94].
2.2 Principles of Elementary Combinatorics
The following principles ofelementary combinatorics
form a significant part of most syllabi in DM:
• Principle of multiplication;
• Principle of addition;
• Principle of bijection;
• Pigeonhole principle;
• Principle of inclusion and exclusion.
Most of these principles are quite simple, some might
even say trivial. However, it is very important to
convince students that when used together, they provide
powerful tools. To be used effectively, students must
perform certain steps to reduce an initial (perhaps quite
sophisticated) problem to a number of very elementary
standard problems.
In this paper, we pay special attention to just two of
these principles: the Principle of Bijection and the
Principle of Inclusion and Exclusion. The Principle of
Bijection may be formulated as follows: For two sets A
and B, we get equality of cardinalities if and only if there
exists a bijection between the sets A and B.
Example 2.1 Count the subsets of the set {1,2,3}.
Initially, we solve this problem by what is called the
“brute force approach.” We list in a certain order all
subsets, and then we list separately all binary sequences
of length 3 as follows.
{} 000
{1} 100
{2} 010
{3} 001
{1,2} 110
{1,3} 101
{2,3} 011
{1,2,3} 111
After this, we discuss the existence of a natural
bijection between the two listed sets which is obtained
by means of characteristic functions for the subsets of
the set {1,2,3}. Evidently, such a bijection can be
established for an n-element set, where n is an arbitrary
natural number. Now, using the Principle of Bijection
in conjunction with the Principle of Multiplication, we
obtain a formula for the cardinality of the power set of
an n-element set, namely 2ⁿ.
Example 2.2 Determine the number of natural divisors of
5! = 120.
We note that 120 =23
·3 ·5. Therefore, for each natural
number d, we see that d divides 120 if and only if d
factors as 2a*3β*5δ with

0 ≤ α ≤3,
0 ≤ β ≤1,
0 ≤ δ ≤1.
Once again, a combination of the Bijection Principle
and the Principle of Multiplication leads to the desired
result: (1+3)(1+1)(1+1) = 4·2·2 = 16.
We now discuss the Principle of Inclusion and Exclusion
(PIE). This principle should be introduced after a careful
consideration of all other principles of elementary
combinatorics, so that students are able to immediately
see its applications, ranging from those that are trivial to
those that are more sophisticated. This provides one more
opportunity to bring together topics from both logic and
set theory, and to examine them from a combinatorial
point of view. To give students an intuitive feeling for
this principle, it is usually a good idea to begin with a
simple example, such as the following.
Example 2.3 Consider a group of 30 students, 18 of
whom have a driver’s license, 20 of whom are native
English speakers, and 7 of whom both have a driver’s
license and are native English speakers. Prove that there
is an inconsistency in this data.
We may model this by a Venn diagram with two
(intersecting) circles, representing the native English
speakers and students with driver’s licenses,
respectively. We let x,a,b,c denote the cardinalities of
the four regions in this Venn diagram, as indicated

Михаил Клин 003


Then one easily obtains the following linear system of
x +a +b +c = 30
a + c = 20
b + c = 18
c = 7
Solving these equations simultaneously shows that
x = -1, an obvious contradiction. After PIE is
introduced, one can revisit this example and obtain the
contradiction immediately from the computation
x = 30 – (20 + 18) + 7 = -1.
Of course, it is important to quickly move on to more
challenging problems where the power of PIE becomes

There are many natural links between this topic and
elementary number theory. In particular, at this stage in
the course, one may introduce the Sieve of Eratosthenes
and Euler’s totient function, φ(n). Depending on the level
of the audience, these linksmay be considered in a
rudimentary form with the aid of a few simple examples,
or in a more rigorous fashion with all the theorems and
2.3 Counting Methods in Elementary Combinatorics
Elementary combinatorics deals with such simple
combinatorial objects as arrangements, combinations, and
permutations. In order to work with a concrete
combinatorial object, one has to consider certain kinds of
representations. In many cases, the same object has more
than one representation. For example, {{a,b}, {c,e},
{d,f}} = {{b,a}, {f,d}, {c,e}} = {{d,f}, {e,c}, {b,a}} = …
are just a few of the many (in fact, 48) distinct
representations for the same partition of the set
{a,b,c,d,e,f}. Thus, it is extremely important to define a
canonical representation for an object. In the partition
example just given, we can agree to start each subset in
the partition with the element appearing earliest in the
alphabet, and then order the subsets in the partition
using dictionary order. For instance, in the example
above, the first representation for the given partition is
canonical in this sense. It would be difficult to overstate
the importance of canonical representations in discrete
mathematics. In order to be able to manipulate
canonical representations, a student should have a
thorough understanding of partially ordered sets, Hasse
diagrams, lexicographic order, and significant exposure
to numerous examples of ordered sets in various
Understanding how, in principle, to define a canonical
representation and how to conduct a brute force search
for all distinct objects is a crucial algorithmic skill. Work
with canonical representations also provides a natural
opportunity to consider equivalence classes and to
observe quotient sets in “real mathematical life”. This is
nothing other than the art of enumeration of simple
combinatorial objects. According to the pedagogical
traditions of the former Soviet school, we distinguish
constructive and analytic enumeration (the term
“constructive enumeration” was coined by I.A. Faradzev
in [Far78]). By constructive enumeration we mean the
creation of a complete list of all objects of a desired type.
Analytic enumeration, in contrast, is simply determining
the cardinality of such a list (see [KliLP96] for a detailed
discussion of this topic).
Classical objects of elementary combinatorics appear
through the consideration of such typical dichotomies as
ordered and unordered objects, and objects with or
without repetition (ordered and unordered sets/multi-sets
in another terminology). One rigorous way to define all
such objects is through the use of functions, f: X→Y. For
each set X (and Y) there are two options for the elements:
distinguishable or indistinguishable. Furthermore, the
function f typically belongs to one of three possible (not
mutually exclusive) classes: injective, surjective or
arbitrary. Thus, consideration of these types of problems

provides 2·2·3 = 12 classical combinatorial problems
which form the kernel of elementary combinatorics.
Such a formal understanding of the essence of
enumeration is surely an achievement of mathematical
didactics. Its roots originated in the 1960’s and 1970’s.
Surprisingly, this strict formulation first appeared, to the
best of our knowledge, in the literature quite recently (see
[Sta86], where in section 1.4 this set of problems is
referred to as the “Twelve-fold Way”).
Example 2.4Determine the number of different 4-letter
“words” which can be constructed from the letters in
In an advanced class, one should explain that this
problem can be solved by using exponential generating
functions. However, even in such a class, we believe
that an initial brute force approach is preferable. This
approach begins with a clever decomposition, briefly
presented in Table 1 below.
Table 1: DecompositionМихаил Клин 004


Using Table 1, it is possible to provide students with a
visual explanation which utilizes a combination of the
Principle of Addition and the Principle of
Multiplication. We also naturally consider ordered and
unordered objects, examine the significance of objects
with repetition, and encounter once again the formulas
for the number of objects of a prescribed type (although
in this example such numbers may be obtained through
a routine listing). Being able to solve this problem
shows a fundamental understanding of the principles of
elementary combinatorics.
2.4 Isomorphism of Graphs
One of the definitive achievements in the evolution of
teaching strategies in DM during the last few decades is
the understanding that the notion of graph isomorphism is
a crucial ingredient in the course. This concept should be
discussed as soon as possible, and definitely should not
be postponed until the end of the course. Consider the
following small selection of modern textbooks in DM:
[Gri94], [Big89], [GooP02], and [Tuc95]. While they
differ on a number of important features, each of these
textbooks considers isomorphism at the beginning of the
chapter(s) devoted to graphs. The reason is quite clear; an
adequate understanding of graphs is practically
impossible without an understanding of isomorphism.
As a rule, when introducing graphs one begins with a
rigorous set-theoretical definition of a graph Γas a pair Γ
= (V,E) consisting of the set V of vertices and the set E of
edges (or arcs). The set E canbe identified with a subset
of Cartesian product V2
(for directed graphs with loops)
or with a subset of the set of two-element subsets of V

(for simple graphs). Initially, it is probably a good idea to
concentrate on the class of simple graphs; that is,
undirected graphs without loops or multiple edges. One
should stress that two graphs are equal by definition if
they are equal as set-theoretical objects: that is, if they
have the same vertex sets and the same edge sets. It is
important to give examples of different representations
for graphs, such as lists (of vertices and edges), diagrams,
adjacency lists, adjacency matrices, and incidence
Constructively enumerating the set of all distinct
graphs with a prescribed number n of vertices and a
prescribed number m of edges turns out to be a nice
exercise in elementary combinatorics, although not all
students immediately realize the essence of the problem.
Indeed, letting n = 4, m = 3, and fixing the vertex set as V
= {1,2,3,4}, we must choose 3 of the 6 two-element
subsets of V for our edge set. Thus, we have 20 distinct
graphs. It is quite useful to depict (or have students
attempt to generate) all such 20 graphs, such as (think of
each graph as having the same labeling on the vertices):


Михаил Клин 005

One then raises the question: which of these graphs are
essentially different and which are essentially the same?
This leads to a rigorous definition of graph isomorphism.
Returning to the example above, one obtains exactly three
isomorphism classes of such graphs. In fact, we can
arrange our list of all 20 graphs so that the isomorphism
classes will be grouped together. This also provides a
natural opportunity to illustrate canonical representations.
Another extremely important concept is that of an
abstract graph, or graph without labels on the vertices.
This is really an isomorphism class of labeled graphs
(usually with a prescribed vertex set, although this is not
obligatory). Some examples of abstract graphs are:

Михаил Клин 005


This topic provides an opportunity to demonstrate to
students that a concrete pictorial view of the diagrams is
not essential for isomorphic graphs (provided that we are
not interested in topologically inequivalent drawings of
the same abstract graph). This is analogous to the
description of the grin of the Cheshire cat in Lewis
Carroll’s Alice in Wonderland.“‘All right,’ said the Cat;
and this time it vanished quite slowly, beginning with the
end of the tail, and ending with the grin, which remained
some time after the rest of it had gone. ‘Well! I’ve often
seen a cat without a grin,’ thought Alice; ‘but a grin
without a cat! It’s the most curious thing I ever saw in all
my life!’”
This naturally leads to a more serious investigation of
the famous graph isomorphism problem. That is, we can
discuss the ambiguity of whether a “good” solution exists
for the problem of determining when two graphs are

isomorphic. For many audiences, this idea may not be
accessible. In advanced classes for mathematics and
computer science majors, the concept of polynomial time
algorithms may be briefly discussed (see [Luk93] for
more details about the status of this problem). However,
for most audiences, this problem will be addressed on a
naïve heuristic level.
We refer to [KliPR88] and [KliRRT99] for more details
concerning educational and computational aspects of the
graph isomorphism problem.

3 Models for DM Courses
The union of the teaching experiences of the authors in
the areas of DM includes several dozen courses delivered
in at least six countries, including the former USSR,
USA, Israel, Fiji, and Germany, although most longstanding activities took place inthe first three countries
mentioned. Based primarily on these experiences, but
also taking into account information provided by our
colleagues, this section is an attempt to represent a
number of different models which may be used in the
teaching of DM. We do not mean to imply that these
models universally cover all possible options for teaching
DM. Nevertheless, they represent quite typical situations
at large universities in the USA, Israel, and other
In our discussion we will refer to various educational
institutions by the following abbreviations:
• UD University of Delaware, Newark, DE;
BGU Ben-Gurion University of the Negev,
Beer Sheva, Israel;
• WIS Weizmann Institute of Sciences,
Rehovot, Israel;
• NTC Negev Technological College,
Beer Sheva, Israel;
• USP The University of the South Pacific,Suva, Fiji;
• FU Free University, Moscow, USSR.
Numerous other institutions will be referred to
anonymously, so as not to overload this presentation with
too many details.
A number of courses that we taught, which definitely
belong to the area of DM, were given to graduate students
(for instance, topics courses on graph theory,
combinatorics, algebraic graph theory, algebraic
combinatorics, finite geometry, and so on [at UD, BGU,
WIS]). Although these were interesting and exciting
teaching experiences, the specific focus and content of
these courses is beyond the scope of this paper.
Nonetheless, it should be pointed out that most of the
pedagogical principles shared by us are valid for all levels
of the audience, from freshmen to final year Ph.D.
In this text, as previously mentioned, by teaching DM
we mean teaching DM to undergraduate students. Several
important factors are considered in our description of the
various models:
(i) Is this a one-semester or two-semester course?
(ii) Are there pre-requisite courses that cover part of the
material in DM? 

(iii) How many contact hours are devoted to lectures and
discussion/exercise classes?
(iv) What is the intended audience for the course (math,
CS, chemistry, engineering, economics, etc.)?
Taking these factors into consideration, we list several
possible models (not necessarily inclusive) for a course in
A. Math and CS Majors – 4 hours of lectures and 2
hours of discussion (exercises), one semester, prerequisite is a one-semester course of mathematical
logic [BGU and FU].
B. All Majors– 3 hours of lectures and 1 hour of
discussion (exercises), one semester, no prerequisites. We call this course DM1 [UD and USP].
C. Communication Engineering Majors– 4 hours of
lectures and 2 hours of discussion (exercises), one
semester, pre-requisite is linear algebra [BGU].
D. Math and CS Majors– 3 hours of lectures and 1
hour of discussion (exercises), one semester, prerequisite is DM1. We call this course DM2 [UD].
E. Software Engineering Majors– 3 hours of lectures
and 1 hour of discussion (exercises), one semester,
no pre-requisites [NTC].
We now very briefly discuss each of these models in
more detail.
A. In this format, we have the greatest number of
teaching possibilities. Studentsenter the class with a
deep understanding of set theory, logic, relations, and
natural numbers. They are familiar with proofs and
the use of formal notation in proofs. The course itself
covers elementary and enumerative combinatorics,
recursion, ordinary and exponential generating
functions, and elements of graph theory. Proofs are
required, including rather sophisticated ones! Many of
the exercises have features of Math Olympiad
problems and require certain non-standard
B & D These two courses form a sequence for math and
computer science majors. Taken together, DM1 and
DM2 cover all the material in Model A (together with
the pre-requisites). This was a typical model during
the first decade of our teaching experience at UD.
Unfortunately, during the last decade, DM2 was not
offered every year, and thus the default DM course for
most BS students became DM1. Another
disadvantage of this model is that the same course is
offered to different audiences, where the range of
abilities and mathematical backgrounds may be
C. This model was created at BGU a few years ago when
the Department of Communication Engineering
realized that DM forms an important ingredient in the
mathematical education of their students. There was a
commitment that elements ofset theory (on a naïve
level) would be considered in Linear Algebra, thus
ensuring that second year students would begin DM
with a reasonable level of mathematical maturity. In
fact, most of the syllabus for model A (as well as
logic, relations, functions, and natural numbers) is
included in the syllabus for model C. In such a course,

although the level of mathematical rigor is clearly
reduced, students still encounter formal proofs and
many sophisticated techniques from enumerative
E. This course is quite similar to model C, although there
are fewer contact hours and the audience is somewhat
weaker. It should also be noted that DM was typically
offered at NTC to this audience in the same semester
as a computer software course which supported some
logical languages. For the majority of the audience
this combination of courses proved to be very
These models will be referred to by name in subsequent
sections as the need arises.
4 Vertical and Horizontal Educational Links
We borrow this term from the NSF program VIGRE
(“Vertical Integration ofGraduate Research and
Education”), although our use of the term is somewhat
different. Briefly, by a vertical educational link, we mean
a prerequisite structure for taking undergraduate courses.
This can be done formally, where certain courses must be
taken before one is allowed to register for a given course,
or informally with a “consent of instructor” prerequisite,
where students registering for a course must first talk to
the instructor to make sure they are indeed prepared for
the material that will be presented in that course. One can
also speak of “degenerate” vertical links within a course,
where mastery of one topic is needed before certain other
topics (or techniques) can be addressed.
By a horizontal educational link, we mean a reference
to some other course, at about the same maturity level,
that a student is taking, has taken, or is about to take.
Such references are often made by instructors to recall
certain ideas or to give students a preview of topics to
4.1 Vertical Links from High School to DM
In teaching DM at the university level, it would be
extremely helpful to know that the elementary principles
of set theory, logic, and basic counting were covered in
the high school curriculum. At one point in time this was
true, at least in a few countries. A good example of this is
the “revolution” in mathematics education at the
secondary schools in the former Soviet Union that took
place in the 1960’s (see [VilS74] for a discussion of this
reform by one of its most important activists, N.Ya.
In the United States, the reform movement in
mathematics education began in the 1950’s with the
formation of the UICSM (University of Illinois
Committee on School Mathematics) in 1951 and the NSF
(National Science Foundation) in 1950. By 1958 after
several revisions, courses for the four high school grades
were implemented in a dozen pilot schools. These
materials were divided into 11 units, which covered all
the topics in the usual secondary school program. The
two facets of understanding central to the development
and the methodology of this curriculum project were the
precision of language and the discovery of
generalizations. The inclusion of DM topics, such as set

terminology and notation, logic, and deductive structure
and theory served the purpose of adding clarification and
precision rather than of making the mathematics more
rigorous. See [Hen63] for a more detailed discussion of
this development.
Today many of these original 11 units form the core of
courses in high school mathematics. However, by the late
1970’s the classroom-piloted and revised materials had
been replaced by commercialtextbooks and much of DM
had evolved from central organizing themes to selfcontained chapters and topics in chapters. These topics
might then be omitted by a teacher because students
found them difficult and parents did not grasp their
mathematical value. Today, some discrete topics continue
to be addressed in the NCTM Curriculum and Evaluation
Standards for School Mathematics, see [NCTM00].
However, an instructor of DM at the university level
simply cannot assume that prospective undergraduate
students have been exposed to these DM concepts in high
school. This is in stark contrast to the experience of a
calculus instructor at the university today, and is one of
the reasons why teaching DM is often considered a more
“challenging” assignment. Thus, for all intents and
purposes, in the United States, the vertical link between
high school mathematics and DM at the university is
tenuous at best.
4.2 Vertical Links from Pre-Calculus and Calculus to
At many universities in the United States today,
significant numbers of students take some sort of precalculus, college mathematics, or contemporary
mathematics course. This is often a requirement for most
students at the university. In a contemporary mathematics
course one often finds an elementary discussion of some
topics from DM. However, this course is not a
prerequisite for subsequent courses.
Unfortunately, most students taking such a course do
not go on and take additional mathematics courses at the
university. Hence, once again, any vertical link between
such courses and DM is tenuous at best, and probably
should be ignored.
However, more and more students are taking calculus
in high school, and most students in DM have had some
exposure to this subject, either in high school or in a
previous university mathematics course. Thus, while
calculus is not a formal prerequisite for DM, most
instructors in DM can safely assume that the vast
majority of their students are familiar with this content
area. This link can be useful in two ways. First, exposure
to calculus should increase a student’s mathematical
maturity. Second, there are some calculus topics that
often appear in DM, such as functions, sequences, and
series. For example, sequencesare typically defined by a
formula in calculus, while theyare most naturally defined
recursively in DM. When series are presented in calculus,
the fundamental issue is one of convergence. In DM,
however, series are typically treated formally as tools for
counting with no concern about convergence issues.
Sometimes these links can effectively be used by
attentive instructors of DM.

4.3 Vertical Links from DM to Other Courses
The previous sections indicate that currently there is at
least one viable link from other courses to DM
(“incoming arcs”). It should also be noted that one
corollary of the relative absence of incoming vertical
links to DM is that probably most universities should
offer a number of different courses in DM, distinguished
by the mathematical background and interests of the
student audiences. However, this is basically a funding
issue and will not be discussed further in this paper.
4.3.1 Vertical Links from DM to Probability
It is widely accepted that probability (and statistics)
should play a significant role in any mathematics
curriculum. There is an obvious vertical link from DM to
probability; namely, counting techniques and basic
enumerations are fundamental to any discussion of
discrete probability. Either one of the two previously
discussed prerequisite structures can be employed here.
One can informally suggest to students that knowledge of
DM will be extremely beneficial in the study of
probability, but not formally require DM as a
prerequisite. In this case, formal topics from probability
theory would not necessarily appear in the DM syllabus,
and historically one would use a book such as [Fel68] for
the probability course. As this book is no longer in print,
one can easily substitute [Gha99] for this task.
Alternatively, one can formally require DM as a
prerequisite for probability. In this case, the syllabus for
DM (see Section 2) should include topics such as the
elementary rules of probability, Bernoulli trials, and
expected value. A possible text for such a course in DM
is [BakE99].
4.3.2 Vertical Links from DM to the Design of Algorithms
About two decades ago when DMwas introduced into the
university mathematics curricula and when computer
science departments had a surplus of students, it was
common practice was to include an introduction to the
analysis of algorithms in the DM syllabus. At that time
CS departments seemed content to let math departments
play the role of introducing this topic to their students.
Thus, program correctness and time complexity analysis
often appeared in DM syllabi and textbooks at that time
(see [BakE99], [Big89]).
Today, however, when CS departments have more
faculty and fewer undergraduate majors than they did a
decade ago, these algorithmic topics have typically
moved back into the CS curriculum. Nonetheless, current
DM courses still include “algorithmic approaches” to
many problems. This not only prepares CS majors for a
formal Design of Algorithms course to be taken later on,
but also gives all students in DM exposure to constructive
proofs and algorithmic thinking. Problems in DM where
this is particularly appropriate include finding gcd’s
(Euclid’s Algorithm), elementary primality testing (Sieve
of Eratosthenes), 2-coloring bipartite graphs, and solving
linear recurrence relations. It should be emphasized that
the actual implementation of algorithms is best left to
computer science courses.

4.3.3 Vertical Links from DM to Graph Theory
Graph Theory has become pervasive in most science and
social science disciplines. Thus the link from DM to
graph theory is a crucial one. How much can be included
in the syllabus depends upon whether DM is a onesemester or two-semester course (see Section 3). In any
case it is important to quickly get beyond the typical
lengthy list of definitions at the start of graph theory to
topics that are mathematicallymore significant. Since the
follow-up courses in graph theory can be either
theoretical or applied, the nature of this vertical link is
Graph algorithms, such as Kruskal or Prim for
minimum cost spanning trees, Dijkstra for shortest paths,
and the Hungarian Method for maximum matchings in
bipartite graphs, form a significant ingredient in this
vertical link. These algorithms will prepare students for
many areas of applications. If time permits, a careful
analysis of at least one of these algorithms, including a
rigorous proof of correctness and a time complexity
analysis, should be done.
4.4.4 Horizontal Links
Horizontal links are the references to other courses and
walks of life that naturally relate to the topic being
discussed in class. It is hoped that all experienced
teachers of DM regularly do this to keep the material
alive and relevant. For instance, an opportunity for this
activity occurs when changing the index of summation in
various enumeration problems. Students should be
reminded that this is analogous to the method of
substitution when computing integrals in a calculus
At some institutions collaboration between the
mathematics department and various “customer
departments” sometimes enables the instructor of DM to
tailor the syllabus to the specific needs of the audience.
This occurred at FU in 1990-92, where the audience
consisted primarily of chemistry majors, and at BGU,
where a version of DM was tailored to the needs of
students studying communication engineering (using the
textbook [Big89]).
5 Further Philosophical and Logistical Facets of DM
In this section we consider teaching DM from a global
perspective. We discuss the overall goals of a DM course,
the skills we want students to achieve by taking DM, the
nature of proof and its relation to DM, the use of exercise
classes and discussion sessions, the selection of
textbooks, and the evaluation of students.
5.1 Key Goals of a Course in DM
First we elaborate the overall goals of DM, at least in our
opinion. In particular, we indicate how these goals might
differ when viewed from the perspective of a CS
department as opposed to a math department.
5.1.1 Paramount Objectives
DM as an independent educational subject was defined
within the last 40-50 years.The year that we might

attribute to its birth would be 1957, the year in which the
first sputnik was launched. This historical event created
great challenges for principally new applications of
mathematics in what had suddenly becoming a growing
technological era (see the discussion in Section 5.4
below). From its early beginnings, DM was intentionally
contrasted with continuous mathematics, and was once
called a “grab bag of methods” by the eminent
mathematician Saunders MacLane in response to an
article promoting DM written by Anthony Ralston
[Ral84]. This new scientific area had at its kernel
combinatorics and graph theory, which in turn was linked
with set theory, logic, algebra, geometry, algorithms, and
optimization. Within the arena of the mathematical
education of students, we suggest that DM has particular
educational goals (perhaps, even a special mission) which
previously had not been addressed in the teaching of
Some of these meta-goals were formulated 20-25 years
ago in the framework of the ACM standards for a
curriculum in undergraduate computer science in the
United States (see [Ber87] and the references in that
article). Below are stated six features, the first two which
are quotes from [Ber87], that we believe are essential to
any course in DM.
• “DM is a terse and precise language of mathematical
• “A primary use of DM is to make difficult problems
• Learning proof techniques is an essential feature of
• DM provides the mathematical background for
algorithmic thinking.
• DM provides an alternate paradigm to continuous
mathematics in mathematical modeling.
• DM is considered by many instructors to be the ideal
breeding ground for creative mathematical thought.
(Perhaps this point of view is best driven home in
[CheK92], where problems borrowed from various
International Mathematical Olympiads form the
backbone of this textbook.)
5.1.2 Math and CS Perspectives
In most universities there is a joint course in DM for both
CS and Math majors. As a rule, the syllabus for such a
course is a compromise between the interests expressed
by these two departments. On the one hand, a course in
Discrete Structures (rather than DM) probably would best
serve the algorithmic interests of most beginning CS
students. Topics in such a course would include Boolean
algebra, Turing machines, formal logic, searching and
sorting, and various programming applications (see
[Lev80] for a typical textbook). On the other hand, DM
as a subject in pure mathematics is a more sophisticated
object. In this context DM may be regarded as a first step
in the mathematical education of someone who will
eventually work in some area of algebra, geometry,
combinatorics, graph theory, probability theory, and so
on. Emphasizing proofs, introducing the notions of
recursion and generating functions, and providing

5.1.3 Achievement Skills
In this section we describe the skills we believe all
students taking DM should achieve. These skills are more
or less independent of the model for DM being offered
(see Section 3), although obviously the course syllabus
will have some impact on what is possible to achieve.
• Students should finish a course in DM with a fluency
in combinatorial reasoning. That is, they should be
able to use the basic principles of elementary
combinatorics, they should be able to decompose
combinatorial problems into a sequence of simpler
problems, and they should be able to manipulate
various representations of the standard combinatorial
• Students in DM should learn algorithmic thinking. In
particular, students should have enough common
sense to try brute force approaches as a first attempt,
should understand divide and conquer strategies, and
should understand how to represent combinatorial
objects on a computer.
• Students should understand the art of combinatorial
counting. This is especially true for math majors, and
includes such skills as the effective use of ordinary
and exponential generating functions as well as the
use of recursion.
• Students should be able to use graphs to model a wide
variety of problems. Of course, time limitations
become a serious concern in developing this skill,
especially in one semester courses. Thus we see once
again the importance of introducing graphs as early as
possible in DM.
• Students should learn the essence of classification.
This includes an understanding of equivalence
relations and isomorphisms. Perhaps this is most
important for future applications both inside and
outside mathematics. That is, students need to learn
how to avoid needless mess and chaos, and how to
create order out of disorder. 

5.2 Proofs
Although the role of proof was previously discussed (see
Section 2.1), in this section the perspective taken is more
global. Clearly, establishing the validity of specific
results is the heart of any mathematical activity. Many of
the proofs in DM are required for the future development
of the subject matter. For example, proofs are required to
establish whether a given relation is an equivalence
relation, to confirm the correctness of an elaborate
definition, to establish a practical criterion for the
fulfillment of some necessity condition, and to
demonstrate that an algorithm will terminate after a finite
number of steps. Moreover, in DM many proofs are

constructive in nature, and thus are useful algorithms in
their own right.
Typically there is not enough time to provide all
possible proofs in a course of DM. One should choose
proofs that potentially increase the students’
understanding of the mathematics involved. In many
cases, a complete proof may be replaced with an outline
of the proof, with a proof of a particular case, or with a
particularly striking example.
When is a proof to be considered “beautiful” or “from
the book”? There is no reasonable way to answer this
question; it is simply a matter of taste. A good example is
a combinatorial proof of a binomial identity, where each
side of the identity is shown to represent the number of
elements in the same set of combinatorial objects
(counted in two different ways). The stronger students in
a class often appreciate the beauty in such an argument,
while the weaker ones typically are befuddled by the
Finally, we mention the role of experimentation.
Mathematical experimentation, followed by conjecture,
and then eventually proof is the standard mode of
operation for most practicing mathematicians. It is always
nice to introduce this practice in the classroom, time
permitting. As the use of computers and symbolic
manipulation packages becomes more and more
prevalent, this mode ofteaching will undoubtedly
become more common. Instructors will have to change
their behavior to allow for more cooperative efforts with
the students inside the classroom. There are many parts of
DM where such experimentation seems particularly
appropriate, elementary number theory being a prime
5.3 Exercises and Discussion Sections
As in other mathematics course, lectures in DM should be
followed by exercises or discussion sections. There are
two main forms for this activity. In most European (and
Israeli) universities, small groups of students will meet
each week for two hours witha teaching assistant (TA,
for short), who may be a graduate student or a faculty
member. These meetings are called exercise classes. The
TA plays an active role in these meetings, illustrating the
lectures of the previous week through some “warm up”
exercises, followed by some standard exercises, and then
some more challenging exercises. The majority of
students taking DM at these institutions are quite serious,
and both the lecturer and the TA have office hours that
are fully utilized by the students.
In most American universities the situation is
somewhat different. Again small groups of students meet
with a TA once per week. But here the meeting time is
usually one hour, and the TA is definitely a graduate
student. Moreover, the role of the TA is much more
passive. That is to say, the TA spends most of his time
answering student questions, serving as a mediator
between the lecturer and the students, and helping
students overcome various difficulties facing them
(sometimes, psychological ones). Office hours for the TA
tend to be more heavily attended than for the lecturer,
partly because the TA again is viewed as an advocate.

5.4 Textbooks
Writing a textbook is undoubtedly the clearest possible
way of expressing one’s teaching philosophy and
methodology concerning a given subject area. We have
supported this thesis in practice, writing two textbooks in
the area of DM ([BakE99] and [KliPR88]).
When discussing textbooks in DM, one needs to start
again in the year 1957 with the pioneering book
Introduction to Finite Mathematics by Kemeny, Snell,
and Thompson [KemST57], whose last edition appeared
in 1974. This was an innovative textbook, not only by its
content but also by its style and method of presenting
material. It is important from a historical point of view to
note that the title calls the subject area “finite
mathematics”, not “discrete mathematics”. Finite
mathematics still exists today, at least in the USA, as a
separate course from DM. It is taught primarily to
business students (and sometimes biology students), and
emphasizes matrix arithmetic, linear programming, a
little finite probability, and perhaps an introduction to
Markov chains. There is little overlap with DM, at least
as defined in this article.
Restricting now to DM as defined here, there are a few
textbooks that we feel deserve special credit due to their
essential input in the development of DM as an
educational subject. In particular, we mention [Vil69],
which lead to the establishment of the Russian tradition
of teaching elementary combinatorics as part of DM, and
the comprehensive book [Liu68], which was the principle
root of all serious teaching literature in DM. We consider
these the forerunners of modern textbooks in DM.
Of the modern textbooks in DM that we have tested,
we feel the following are among the best: [Bal91],
[Big89], [Bru04], [CheK92], [GooP02], [Gri94],
[Rob84], and [Tuc95]. We also mention a few classics;
these are books not necessarily intended for
undergraduate students, but rather are encyclopedic
sources of information for instructors of DM. These
include [GraKP94] and [Sta86]. Assuming a
homogeneous audience, we believe that currently there is
no problem selecting a good textbook (or textbooks) for a
course in DM. For example, we used [Tuc95] for Model
B, [Big89] and [Gri94] for Model C, and [Bal91] and
[Gri94] for Model E. Each time the audience seemed
quite happy with the text selection. Moreover, it is worth
mentioning that the supply of textbooks in DM is
exploding with each passing year. For example,
depending upon one’s taste and the nature of the
(homogeneous) audience, anyone of the recent texts
[Ros03], [LovPV03], [PemS03], [Wal03], [Epp04], or
[WheB05] could be effectively used.
However, for the advanced students in Model A or for
a non-homogeneous audience in Model B, there is still a
serious problem. There does notseem to be any available
text written in a multi-level fashion that can
simultaneously be used by a range of students, varying
from fairly weak to exceptionally strong. This is
especially important for large American universities with
diverse audiences, where tradition and economics dictate
that only one textbook be assigned. One of the goals of
this paper is to inspire others to write such a definitive

5.5 Grading and Exams
Strong differences in the culture, traditions, and types of
courses offered at various universities around the world
make it impossible to make any universally accepted
statements concerning the giving of exams and the
awarding of grades in DM. However, we list below a few
models with which we have some experience, and then
describe some of salient features of each.
5.5.1 The Soviet Pattern
Most of the exams in the former Soviet Union (and many
other European countries, as well) were oral. Typically, a
lecturer prepares a few dozen “tickets”, each of which
contains one or two theoretical questions and one or two
exercises. The theoretical questions usually involve the
formulation and proof of some standard proposition.
Students then randomly select a ticket and are given 30-60 minutes to prepare an answer, after which they orally
present their solution in 15-20 minutes before the
examiner (instructor and/or TA). Students might be given
extra time to clarify some claims or provide an answer to
some supplementary question.The exam is then graded
on a basis of 5 to 2 (in the former Soviet Union style),
which is roughly equivalent to the A to D basis in the
U.S. Sometimes plus/minus grading is used.
The main advantage of such a system is the fact that it
allows for almost all of the presented material to be fair
game for the exam, including most of the proofs given in
the course. This partly explains why the level of
mathematical rigor among university students was so
high in the former Soviet Union. Namely, this exam
pattern pushed students to become comfortable with the
theoretical mode of thinking in all their courses, including
calculus. As a byproduct, DM did not have to concentrate
on introducing proof and abstraction; this was already
accomplished before students took DM. One
disadvantage of this system is the obvious element of
subjectivity in the awarding of marks.
5.5.2 The Israeli Pattern
In Israel course grades are primarily determined by the
final exam, which lasts at least three hours. Sometimes
there is one midterm exam, which may account for 20-25% of the final grade. Homework assignments may
account for an additional 5-10% of the course grade.
Nonetheless, the final exam is by far the dominant factor
in grading. Students are allowed to take the final exam
twice. If a second attempt is made, the results from the
first attempt are automatically cancelled, even if the first
score is higher. This gamble often causes great anxiety
among the students. Each instructor finds his own way of
preparing students for the final.
Psychological support is needed, encouraging students
to do their very best on the first attempt at the final, and
yet preparing them for the possibility of failure. Testtaking skills should be discussed, such as telling each
student to find the easiest problem and do it first. This
tends to calm the nerves and also maximize the total
score. Which problem is easiest depends upon the relative
strengths of the student. Those who like to think

algorithmically and are good at computation will choose
one kind of problem, while those with strong creative
abilities and a distaste for messy computations will
undoubtedly make a very different choice.
It should be mentioned that sometimes choice is
allowed on the final exam, say by allowing students to
choose 5 of 7 problems. This eliminates some of the
pressure, and allows the instructor to better find out what
the students know.
5.5.3 The American Pattern
The typical American patternfor awarding grades puts
less emphasis on the final exam and places more
emphasis on homework, quizzes, and (several)
intermediate exams. Most courses have a total number of
points, say 600, that can be awarded to each student. The
final letter grade for the course (A through F, with
possible plus/minus grading) is determined by the
number of points earned. This correspondence can vary
dramatically from one course to another, and even from
one instructor to another within a given course. However,
in multi-section courses at some universities, there is a
course coordinator and sometimes common exams (at
least, a common final) are given. In such cases, the
correspondence between accumulated points and final
letter grades is often spelled out by the coordinator, with
only minor variations available to individual instructors.
In the American system the predicted final grade
becomes clearer and clearer to individual students with
each passing stage of the course. The final exam
primarily serves as a confirmation that the anticipated
grade is indeed the correct one. Student complaints about
final grades are relatively rare in this system, at least in
comparison to the Israeli system. Students do not feel a
great deal of pressure during final exams, and pretty
much know what grade they will receive in each of their
courses. This certainly can be viewed as an advantage of
the system.
The main disadvantage is the relatively short time
period for each exam. Midterm exams typically last only
50 minutes. Final exams are most often two hours in
length. In a course such as DM this is not enough time to
adequately attack complicated theoretical or
computational problems, especially on the midterms.
Hence, exams tend to test student knowledge somewhat
systematically, but not very deeply.
5.5.4 A Sample of Problems
Below we present a handful of problems which are taken
from actual DM exams given by us and which are
somewhat different than the standard ones found in most
textbooks. Of course, the audience and type of course
determines whether or not such problems are appropriate.
• Let A={1,2,3,4,5,6,7,8} and let R be an equivalence
relation defined on A. Suppose we know that
a) (1,2), (2,3), and (7,8) are in R, while (1,5) and
(1,6) are not in R;
b) |R|= 22 (that is, R consists of 22 pairs).
Determine all possibilities for the quotient set A/R.
• In how many ways is it possible to select three
distinct numbers from {1, 2, 3,…, 99} whose sum is
congruent to 0 modulo 3? 

• Find a closed formula for the sum from k=0 to k=n of
• How many different 7-digit (decimal) numbers
contain exactly 4 distinct digits?
• Find the number of spanning trees in the following


Михаил Клин 011


6 Concluding Remarks
Discrete and continuous mathematics are objectively
distinct parts of mathematics, and this fact is now well
recognized by the mathematical community. Nonetheless,
the role of DM in many math departments is still
undergoing change. This should not be done in a
revolutionary manner. Changes should occur in stages,
taking into account the existing traditions and goals in
making the change.
The success of any model for a course in DM depends
heavily upon the syllabus and the textbook. Teaching DM
as a two-semester sequence is definitely preferable, but
not always feasible. When teaching only a one- semester
course, it is better to decrease the amount of new material
covered and spend the extra time making sure the
audience has the required background knowledge for the
course. In any case, even when the audience is quite
mathematically mature, completing a course in DM will
likely leave the students with a broad spectrum of
emotional experiences, from inspirational highs achieved
through steady progress in combinatorially and
algorithmic thinking to the depressing lows caused by the
necessity of learning this new language of mathematics.
Thanks goes to G. Tinhofer and P. Gritzman, who invited
the third author to publicly present some to the ideas in
this article at the Technical University of Munich in
January of 2003. All three authors would like to thank
many colleagues at various institutions, too numerous to
mention, for years of thoughtful, insightful, and
stimulating conversations that eventually lead to the
pedagogical ideas expressed in this article. We thank
Frank Fiedler for his help withthe graphics portion of the
text. We are grateful to I. Anderson, K. Reiss and two
anonymous referees, whose constructive criticism
enabled us to revise the initial version of this paper.
Finally, it should be mentioned that this paper was
written while the third author was visiting the Department
of Mathematical Sciences at the University of Delaware,
and we wish to thank that institution for its support
during this project.
[BakE99] Baker, R.D., Ebert, G.L. (1999). Discrete
mathematics. Dubuque, Iowa: Kendall/Hunt.
[Bal91] Balakrishnan V.K. (1991). Introductory discrete
mathematics. Mineola, NY: Dover.

[Ber87] Berztiss A. (1987). A mathematically focused
curriculum for computer science. Communications of the
ACM, 30 (5), 356-365.
[Big89] Biggs N.L. (1989). Discrete mathematics, 2
New York, NY: Oxford University Press.
[Bru04] Brualdi R.A. (2004). Introductory combinatorics, 4
edition. Upper Saddle, NJ: Prentice Hall.
[CheK92] Chen, C.C., Koh, K.M. (1992). Principles and
techniques in combinatorics. River Edge, NJ: World
[Cuo03] Cuoco, A. (2003). Teaching mathematics in the United
States. Notices of the American Mathematical Society, 50(7),
[Epp04] Epp, S. (2004). Discrete mathematics with
applications. Belmont, CA: Brooks/Cole.
[Far78] Faradzev, I. A. (1978). Constructive enumeration of
combinatorial objects. In: Problèmes combinatoires et thèorie
des graphes (Colloq. Internat. CNRS, Univ. Orsay, Orsay,
1976), Colloq. Internat. CNRS, 260, 131-135.
[Fel68] Feller, W. (1968). An introduction to probability theory
and its applications, Volume 1 (3
edition). New YorkLondon-Sydney: John Wiley & Sons.
[Gar82] Gardner, M. (1982). Logic machines and diagrams.
Chicago, IL: University of Chicago Press.
[Gha99] Ghahramani, S. (1999). Fundamentals of probability,
edition. Upper Saddle River, NJ: Prentice Hall.
[GooP02] Goodaire, E.G., Parmenter, M.M. (2002). Discrete
mathematics with graph theory (2
edition). Upper Saddle
River, NJ: Prentice Hall.
[GraKP94] Graham, R.L., Knuth, D. E., Patashnik, O. (1994).
Concrete mathematics: A foundation for computer science.
Reading, MA: Addison Wesley.
[Gri94] Grimaldi, R.P. (1994). Discrete and combinatorial
mathematics: An applied introduction (3
edition). Reading,
PA: Addison-Wesley.
[Gru84a] Grunbaum, B. (1984). The construction of Venn
diagrams. College Mathematics Journal 15(3), 238-247.
[Gru84b] Grunbaum, B. (1984). On Venn diagrams and
counting of regions. College Mathematics Journal 15(5),
[Hen63] Henderson, K.B. (1963). Mathematics. In Using
Current Curriculum Developments, Washington, D.C.:
Association for Supervision and Curriculum Development,
[KemST57] Kemeny, J.G., Snell,J.L., Thompson, G.L. (1957).
Introduction to finite mathematics. Englewood Cliffs, N.J.:
[KliLP96] Klin, M., Liskovets, V., Pöschel, R. (1996).
Analytical enumeration of circulant graphs with primesquared number of vertices. Sèm. Lothar. Combin., 36, Art.
B36d, approx. 36 pp. (electronic).
[KliPR88] Klin, M.Ch., Pöschel, R., Rosenbaum, K. (1988).
Angewandte Algebra für Mathematiker und Informatiker.
(German) [Applied algebra for mathematicians and
information scientists] Einführung in gruppentheoretischkombinatorische Methoden. [Introduction to grouptheoretical combinatorial methods] Braunschweig: Friedr.
Vieweg & Sohn.
[KliRRT99] Klin, M., Rücker, Ch., Rücker, G., Tinhofer, G.
(1999). Algebraic combinatorics in mathematical chemistry.
Methods and algorithms. I. Permutation groups and coherent
(cellular) algebras. Match, 40, 7-138.
[Lev 80] Levy, L.S. Discrete structures of computer science.
(1980). New York-Chichester-Brisbane: John Wiley & Sons.
[Liu68] Liu, C.L. (1968). Introduction to combinatorial
mathematics. New York-Toronto: McGraw-Hill Book Co.
[LovPV03] Lovasz, L., Pelikan, J., Vesztergombi, K. (2003).
Discrete mathematics: elementary and beyond. New York,
NY: Springer.

[Luk93] Luks, E.M. (1993). Permutation groups and
polynomial-time computation. In: Groupsand computation
(New Brunswick, NJ, 1991), DIMACS Ser. Discrete Math.
Theoret. Comput. Sci., 11, (pp. 139-175). Providence, RI:
Amer. Math. Soc.
[NCTM70] National Council of Teachers of Mathematics (ed.)
(1970). A History of Mathematics Education in the United
States and Canada. Reston, VA: NCTM.
[NCTM00] National Council of Teachers of Mathematics (ed.)
(2000). Principles and Standards for School Mathematics.
Reston, VA: NCTM.
[PemS03] Pemmaraju, S., Skiena, S. (2003). Computational
discrete mathematics. Cambridge, England: Cambridge
University Press.
[Ral84] Ralston, A. (1984). Willdiscrete mathematics surpass
calculus in importance? College Math. J. 15 (5), 371-382.
[Rob84] Roberts, F.S. (1984). Applied combinatorics.
Englewood Cliffs,NJ:Prentice Hall.
[Ros03] Rosen, K. (2003). Discrete mathematics and its
applications. New York, NY: McGraw-Hill.
[Shi94] Shin, S.-J. (1994). The logical status of diagrams.
Cambridge: Cambridge University Press.
[SieK98] Mathematics education as a research domain: a search
for identity. Book 1, 2. An ICMI study. Edited by Anna
Sierpinska and Jeremy Kilpatrick. (1998). New ICMI Studies
Series, 4. Dordrecht: Kluwer Academic Publishers.
[Sta86] Stanley, R.P. (1997). Enumerative combinatorics. Vol.
1. With a foreword by Gian-Carlo Rota. Corrected reprint of
the 1986 original. Cambridge Studies in Advanced
Mathematics, 49. Cambridge: Cambridge University Press.
[Tuc95] Tucker, A. (1995). Applied combinatorics. 3
New York: John Wiley & Sons, Inc.
[Vil69] Vilenkin N.Ja. (1969). Комбинаторика. (Russian)
[Combinatorics] Moscow: Izdat. “Nauka». Vilenkin N.Ya.
Combinatorics. Translated from the Russian by A. Shenitzer
and S. Shenitzer (1971). New York-London: Academic Press.
[VilS74] Vilenkin, N.Ja., Šreder Ju.A. (1974). The concepts
of mathematics and the objects of science. (Russian) Voprosy
Filos., no. 2, 16-26.
[Wal03] Wallis, W.D. (2003). A beginner’s guide to discrete
mathematics. Boston,MA: Birkhauser.
[WheB05] Wheeler, E., Brawner, J. (2005). Discrete
mathematics for teachers. New York, NY: Houghton Mifflin.
Christine Ebert, Prof., Department of Mathematical Sciences
University of Delaware, 510 Ewing Hall, Newark, Delaware
19716, USA
Gary Ebert, Prof., Department of Mathematical Sciences,
University of Delaware, 510 Ewing Hall, Newark, Delaware
19716, USA
Mikhail Klin, Prof., Departmentof Mathematics, Ben Gurion
University,Beer Sheva 84105, Israel



Об авторе

Доктор физико-математических наук


Лимит времени истёк. Пожалуйста, перезагрузите CAPTCHA.