Engineering, Built Environment and IT
Department of Computer Science
Programming Languages
COS 333
Examination
25 June 2024
Examiner: Mr W. S. van Heerden
External Examiner: Mr Willem H. K. Bester (Stellenbosch)
Instructions:
Read the questions carefully and answer all questions.
This section comprises of 28 questions, and a total of 40 marks.
You have 3 hours to complete this test.
This test is an online assessment:
The test will automatically submit when the time (3 hours) expires. You can also submit the test
yourself if you are done ahead of time.
Be sure to complete the assignment submissions of the two practical implementation questions
(Questions 5 and 6). If you do not complete the submission of your implementations, there is no way
it can be accessed for assessment. There are unlimited upload opportunities for the implementation
assignments, but only the last submission will be assessed. Please make absolutely sure that you
upload the correct file for each assignment.
This test is closed book and is subject to the University of Pretoria integrity statement provided
below.
You are not allowed to consult any sources other than the Practical 2 specification, the MIT/GNU
Scheme Reference Manual (provided as documentation for Practical 2), the Practical 3
specification, the SWI-Prolog Reference Manual (provided as documentation for Practical 3), and
the Practical 4 specification.
Your Scheme submission for Question 5 must be working Scheme code that will successfully
interpret using version 8.8 of the DrRacket interpreter. You may (and should) use this interpreter to
test your submission. Once you have finished writing your program, make sure it is saved in a text
file with the appropriate name (described in the question), and submit it to the correct upload slot.
Your Prolog submission for Question 6 must be working Prolog code that will successfully interpret
using version 9.0.4 of the SWI-Prolog interpreter. You may (and should) use this interpreter to test
your submission. Once you have finished writing your program, make sure it is saved in a text file
with the appropriate name (described in the question), and submit it to the correct upload slot.
You may use a non-programmable calculator.
You may not use any other programmable devices.
Integrity Statement:
The University of Pretoria commits itself to producing academic work of integrity. I affirm that I am
aware of and have read the Rules and Policies of the University, more specifically the Disciplinary
Procedure and the Tests and Examinations Rules, which prohibit any unethical, dishonest or
improper conduct during tests, assignments, examinations and/or any other forms of assessment. I
am aware that no student or any other person may assist or attempt to assist another student, or
obtain help, or attempt to obtain help from another student or any other person during tests,
assessments, assignments, examinations and/or any other forms of assessment.
QUESTION 1
1. There are often trade-offs between the programming language evaluation criteria we have used in
this module. Consider the following statements, and select the one that most correctly describes a
likely tradeoff when readability improves.
The cost of execution becomes worse.
The cost of maintenance becomes worse.
The cost of implementation becomes worse.
Reliability becomes worse.
1 points
QUESTION 2
1. An additional programming language evaluation criterion that is not focussed on heavily in this
course is generality. Consider the following programming languages, and select the one that has
the best generality.
Fortran I
ALGOL 60
SIMULA 67
PL/I
1 points
QUESTION 3
1. Consider the following statements about the Pascal and C programming languages, and select the
one that is the most correct.
Pascal is more readable than C.
Pascal is more writable than C.
Pascal is less reliable than C.
Pascal has a higher cost of maintenance than C.
1 points
QUESTION 4
1. An additional programming language evaluation criterion that is not focussed on heavily in this
course is portability. Consider the following programming languages, and select the one that is the
most portable.
Fortran 90
C++
PL/I
JavaScript
1 points
QUESTION 5
1. Write a Scheme function named removePositiveOddValues, which receives one parameter.
The parameter is a simple list containing only numeric atoms. The function should yield (not print
out) a list containing the values in the parameter list, in their original order, with all the positive odd
values removed.
For example, the function application
(removePositiveOddValues '(3 9))
should yield an empty list because all the numeric atoms in the parameter list are positive odd values.
As another example, the function application
(removePositiveOddValues '(-5 3 0 5 9))
should yield the list (-5 0) because 3, 5, and 9 are positive odd numbers, and are therefore
removed from the list, while -5 and 0 remain.
Submit your program code in a file named s99999999.scm (where 99999999 is your student
number) to the assignment submission slot labelled "Examination: Scheme Function", and select the
"Yes" answer below once you have done so. If you do not make a submission, select the "No" answer
below.
Take careful note of the following requirements:
Your submission must be working Scheme code that will be successfully interpreted by version 8.8
of the DrRacket interpreter using the sicp language collections (this means, amongst other things,
that lowercase letters should be used for all built-in function names). You may (and should) use the
DrRacket interpreter to test your submission.
You may define additional helper functions.
Your Scheme function may only use the following built-in functions in the way they are used in the
slides and textbook (failure to do so will result in a zero mark for this question):
o Function construction: lambda, define
o Binding: let
o Arithmetic: +, -, *, /, abs, sqrt, remainder, modulo, min, max
o Boolean values: #t, #f
o Equality predicates: =, >, <, >=, <=, even?, odd?, zero?, negative?, eqv?, eq?
o Logical predicates: and, or, not
o List predicates: list?, null?
o Conditionals: if, cond, else
o Quoting: quote, '
o List manipulation: list, car, cdr, cons
o Input and output: display, printf, newline, read
Yes
No
5 points
QUESTION 6
1. Write a Prolog proposition called doublePositives(L1, L2), which succeeds if L2 is a list
containing all the elements in L1, in their original order, with every positive value doubled in
magnitude, and every non-positive value unchanged. You may assume that lists L1 and L2 are
simple lists containing only integers. For example, the query:
?- doublePositives([-4, -9], X).
should be true with the instantiation X = [-4, -9]. This is because there are no elements that are
positive in the list [-4, -9]. Similarly, the query
?- doublePositives([-3, 6, 7, -1, -5], X).
should be true with the instantiation X = [-3, 12, 14, -1, -5]. This is because -3, -1, and -
5 are not positive and are therefore not doubled, while 6 is doubled to give the value 12 and 7 is
doubled to give the value 14.
Hint 1: You can use comparison operators as terms in Prolog. These operators are similar to what
you are used to in imperative languages. The operators < (less than), =< (less than or equal
to), > (greater than), and >= (greater than or equal to) are all valid. As a very simple example, you
could use a comparison operator as follows:
lessThan(X, Y) :- X < Y.
The following queries are then possible:
?- lessThan(1, 2).
true.
?- lessThan(2, 1).
false.
Note that this is just an example, and does not mean you should use this lessThan proposition in
your implementation (although you may if you wish to).
Hint 2: Multiplication is represented by * in Prolog.
Submit your program code in a file named s99999999.pl (where 99999999 is your student
number) to the assignment submission slot labelled "Examination: Prolog Proposition", and select the
"Yes" answer below once you have done so. If you do not make a submission, select the "No" answer
below.
Take careful note of the following requirements:
Your submission must be working Prolog code that will be successfully interpreted by the SWI-
Prolog interpreter installed on the Windows computers in the Informatorium. You may (and should)
use this SWI-Prolog interpreter to test your submission. For notes on using the interpreter, and re-
querying, see the specification for Practical 3. Once you have finished writing your program, save it
in a text file with the appropriate name, and submit it to the correct upload slot.
You may define additional helper propositions.
Note that you may only use the simple constants, variables, list manipulation methods,
arithmetic and relational expressions, and built-in predicates discussed in the textbook,
slides, and hints. In particular, do not use if-then, if-then-else, and similar constructs. You may
NOT use any more complex predicates provided by the Prolog system itself. In other words, you
must write all your own propositions. Failure to observe this rule will result in all marks for this
question being forfeited.
Yes
No
5 points
QUESTION 7
1. Binding can occur at different times. Consider the following options, and select the one that most
correctly describes the binding time of a standard 32-bit signed integer type in a statically typed
programming language to either a sign-and-magnitude or two's complement representation.
Language design time.
Language implementation time.
Compile time.
Load time.
Runtime.
1 points
QUESTION 8
1. Consider the following program code in C# (line numbers are provided for reference purposes):
1. {
2. int y;
3. {
4. int x = 5;
5. }
6. int x = 2;
7. }
Consider the following statements about the above program code, and select the one that is the most
correct.
The provided program code does not compile because the variable y is not initialised.
The provided program code does not compile because the variable x on line 4 hides the
variable x on line 6.
The provided program code does not compile because the variable x on line 6 hides the
variable x on line 4.
The provided program code compiles and runs successfully.
1 points
QUESTION 9
1. Consider the following statements about character string types in programming languages,
and select the one that is the most correct.
In programming languages that represent character strings using only classes, character strings
are always mutable.
In programming languages that represent character strings using only classes, character strings
are always immutable.
Limited dynamic length strings have a higher overall hardware resource cost than static length
strings, and a lower overall hardware resource cost than dynamic length strings.
C supports character strings as a primitive type.
1 points
QUESTION 10
1. Consider the following statements about unions, and select the one that is the most correct.
Unions are closely related to classes in object-oriented programming, but are more limited in
terms of functionality.
Discriminated unions are more memory efficient than free unions.
Unions can be used as a mechanism for returning multiple values from a subprogram.
Unions are used when memory cost is of great concern.
1 points
QUESTION 11
1. Consider the following statements about arithmetic operations in Scheme, and select the one that is
the most correct.
The order in which arithmetic operations are evaluated in Scheme is unambiguously determined
based on clear precedence levels and associativity rules.
The order in which arithmetic operations are evaluated in Scheme is unambiguously determined
based on function and parameter evaluation.
The order in which arithmetic operations are evaluated in Scheme is unambiguously determined
because all operators have the same precedence level, and associate from left to right.
The order in which arithmetic operations are evaluated in Scheme is ambiguous because
function parameters are evaluated in no particular order.
1 points
QUESTION 12
1. Consider the following statements about expressions in Ada, and select the one that is the most
correct.
The lack of coercions in Ada makes the language more readable.
The lack of coercions in Ada makes the language less reliable.
The common use of coercions in Ada makes the language more writable.
The common use of coercions in Ada makes the language less reliable.
1 points
QUESTION 13
1. Consider the following program code in a hypothetical programming language:
int count = 5
double result = 0.0
while ((count > 0) and (result = (10.0 / count))) {
print result
count = count - 1
}
Assume that the programming language uses C-like syntax, implements assignment as an
expression, uses and to represent a logical and operation, and interprets integers as Boolean values
in the same way that C does. Also assume that the program is syntactically correct. Consider the
following statements, and select the one that is the most correct.
The provided program code produces an error regardless of whether the logical and is short-
circuit evaluated or not.
The provided program code produces an error if the logical and is short-circuit evaluated.
The provided program code produces an error if the logical and is not short-circuit evaluated.
The provided program code produces no error regardless of whether the logical and is short-
circuit evaluated or not.
1 points
QUESTION 14
1. Consider the following statements about unary assignment operators, and select the one that is the
most correct.
Unary assignment operators reduce readability, if they are overloaded to provide prefix and
postfix versions.
Unary assignment operators reduce writability, if they are overloaded to provide prefix and
postfix versions.
Unary assignment operators reduce the execution cost of a programming language.
Unary assignment operators are essential within a programming language.
1 points
QUESTION 15
1. Consider the following statements, and select only those that are true about counter-controlled
loops. Note that incorrectly selected options will be penalised.
Counter-controlled loops are required in order to implement any flowchart-represented algorithm.
Counter-controlled loops in Python are more reliable than counter-controlled loops in C.
Counter-controlled loops in C are more readable than counter-controlled loops in Java.
Counter-controlled loops are not directly supported in purely functional programming languages.
2 points
QUESTION 16
1. Consider the following statements about disambiguating nested selection statements,
and select only the ones that are correct. Note that incorrectly selected options will be penalised.
Python's approach to disambiguating nested selection statements is less readable than Java's
approach.
Python's approach to disambiguating nested selection statements is less reliable than Java's
approach.
Ruby and Perl use very similar approaches to disambiguating nested selection statements, but
Perl's approach is more writable than Ruby's approach.
Python and Ruby use very similar approaches to disambiguating nested selection statements,
but Python's approach is more writable than Ruby's approach.
2 points
QUESTION 17
1. Consider the following program code in a hypothetical programming language that supports guarded
commands:
var myVal = 2
do (myVal < 0) -> print("A ")
myVal = 1
[] (myVal > 0) -> print("B ")
myVal = myVal - 1
od
Assume that the print subprogram prints its string output to the screen on a single line. Provide the
output of the program in the space provided. If the program produces no output, type "Nothing"
without the quotation marks. If the program produces a compile-time error, type "Compilation error"
without the quotation marks. If the program produces a runtime error, type "Runtime error", without
quotation marks, at the point where the runtime error occurs. Note that there is a single space at the
end of each of the outputs produced by the print subprograms.
BB
1 points
QUESTION 18
1. Consider the following statements about subprograms that are defined outside a class definition,
and select the one that is the most correct.
Ruby subprograms that are defined outside a class definition are more writable than C++
subprograms that are defined outside a class definition.
Ruby subprograms that are defined outside a class definition are less writable than C++
subprograms that are defined outside a class definition.
Ruby subprograms that are defined outside a class definition are more orthogonal than C++
subprograms that are defined outside a class definition.
Ruby subprograms that are defined outside a class definition behave in exactly the same way as
C++ subprograms that are defined outside a class definition.
1 points
QUESTION 19
1. Consider the following statements about procedures and functions, and select the one that is the
most correct.
A procedure can produce a result if all its parameters use pass-by-value semantics.
A procedure is guaranteed to produce no side effect if it produces a result.
A function is guaranteed to produce a side effect if it produces a result.
A procedure can more easily produce multiple results than a function can.
1 points
QUESTION 20
1. Consider the following program code in a hypothetical programming language:
doSomething(int first, int second) {
first = 2
print(first)
print(" ")
print(second)
}
void main() {
int val1 = 5
int val2 = val1 + 1
foo(val1, val2)
}
Assume that the print subprogram prints its output to the screen on a single line. Provide the
output of the program code if it is assumed that the programming language uses pass-by-name for
all parameter passing. Provide only the output as it would be printed to the screen, without any
additional text (i.e. additional spaces, quotation marks, or other punctuation). If the program produces
an error of any kind, type "Error" without the quotation marks. Note that the
second print subprogram call outputs a single space.
26
1 points
QUESTION 21
1. Consider the following program code in a hypothetical programming language:
var A = 2
main() {
var A = 5
call sub1(sub2)
}
sub1(param) {
var A = 3
call param()
}
sub2() {
print(A)
}
Assume the programming language allows subprograms to be passed as parameters, that
the print subprogram prints its output to the screen on a single line, and that execution starts with
the main subprogram. Provide the output of the program code if deep binding is used to determine
the referencing environment of the passed subprogram, but the programming language uses dynamic
scoping rules. Provide only the output as it would be printed to the screen, without any additional text
(i.e. additional spaces, quotation marks, or other punctuation). If the program produces an error of any
kind, type "Error" without the quotation marks.
5
1 points
QUESTION 22
1. Consider the following program code in C#:
public static void doSomething<T>(T p1, T p2) {
Console.WriteLine(p1 + " " + p2);
}
Assume that the doSomething method is part of the Program class. Also assume that the following
program code is provided in the main method of the Program class:
Program.doSomething(5, 2);
Program.doSomething(false, true);
Consider the following statements about the use of the doSomething method, and select the one
that is the most correct.
One version of the doSomething method is created at compile time, and is used
with int and bool parameters.
Two versions of the doSomething method are created at compile time. One is used
with int parameters, while the other is used with bool parameters.
One version of the doSomething method is created at runtime, and is used
with Object parameters.
Two versions of the doSomething method are created at runtime. One is used
with int parameters, while the other is used with bool parameters.
1 points
QUESTION 23
1. Consider the following program code in a hypothetical programming language that supports
coroutines:
main() {
resume first
}
couroutine first {
print("X ")
resume second
print("Y ")
}
coroutine second() {
print("A ")
resume first
print("B ")
}
Assume that the print subprogram prints its string output to the screen on a single line, and that
execution starts with the main subprogram. Provide the output of the program code. Provide only
the output as it would be printed to the screen, without any additional text (i.e. additional spaces,
quotation marks, or other punctuation). If the program produces an error of any kind, type "Error"
without the quotation marks. Note that there is a single space at the end of each of the outputs
produced by the print subprograms.
XAYB
1 points
QUESTION 24
1. Abstract data types (ADTs) aim to hide the representations of ADT objects from from the clients that
use them. Consider the following possible programming language features in C++, and select only
the ones that are correct and detract from the effectiveness of ADT representation hiding in C++.
Note that incorrectly selected options will be penalised.
The existence of friend functions and friend classes.
The fact that C++ supports stand-alone functions that are not part of a class.
The fact that C++ requires that classes and member function definitions must appear in the
same file.
The fact that all class data members must appear in header files.
2 points
QUESTION 25
1. Consider the following statements about abstract data types (ADTs) in Ruby, and select the one
that is the most correct.
Instance and class variable names are more readable in Ruby than they are in Java.
Ruby supports multiple constructors, each of which can be implicitly called when new is used
with the appropriate class name and parameters.
It is impossible to implement the functionality of multiple constructors in Ruby.
The interface of a Ruby ADT is very reliable.
1 points
QUESTION 26
1. Imagine a hypothetical programming language in which a subclass can inherit a public method from
a parent class, and change its access control to private or protected. Consider the following
statements about the hypothetical programming language, and select the one that is the most
correct.
In the hypothetical programming language, dynamic binding is less efficient.
In the hypothetical programming language, dynamic binding is impossible.
In the hypothetical programming language, subclasses are subtypes.
In the hypothetical programming language, subclasses are not subtypes.
1 points
QUESTION 27
1. Consider the following options, and select only the approaches that makes Smalltalk more
orthogonal than Java. Note that incorrectly selected options will be penalised.
The difference in approach to the exclusivity of objects in Smalltalk and Java.
The difference in approach to dynamic and static binding in Smalltalk and Java.
The difference in approach to subclasses and subtypes in Smalltalk and Java.
The difference in approach to object deallocation in Smalltalk and Java.
2 points
QUESTION 28
1. Consider the following statements about object-oriented programming in C++ and C#,
and select the one that is the most correct.
C++ provides worse support for multiple inheritance than C# does.
C++ provides less writability in terms of nested classes than C# does.
C++ uses a more orthogonal approach to object-oriented programming than C# does.
C++ provides a more readable approach to dynamic message binding than C# does.
