Carnegie Mellon University

Research Showcase @ CMU

Computer Science Department School of Computer Science

1971

Introduction to ALGOL 60 for those who have

used other programming languages
A. Nico Habermann
Carnegie Mellon University

Follow this and additional works at: http://repository.cmu.edu/compsci

Recommended Citation
.

This Technical Report is brought to you for free and open access by the School of Computer Science at Research Showcase @ CMU. It has been
accepted for inclusion in Computer Science Department by an authorized administrator of Research Showcase @ CMU. For more information, please
contact [email protected].
NOTICE WARNING CONCERNING COPYRIGHT RESTRICTIONS:
The copyright law o f the United States (title 17, U.S. Code) governs the making
o f photocopies or other reproductions o f copyrighted material. Any copying o f this
document without permission of its author may be prohibited by law.
 Introduction to ALGOL 60
for those who have used
other programming languages

by A. N. Habermann

September, 1971

Carnegie-Mellon University
Department of Computer Science
Pittsburgh, Pa. 15213
 Introduction to ALGOL 60
for those who have used other programming languages

by A.N. Habermann

Including Problems and the Revised Report

0. Introduction

1. Data structures

2« Control statements

3. Block structure

4. Procedures

5. Recursion

6. Call by name/value
 -1-

0. Introduction

The idea to write an introduction to ALGOL 60 for people who

have programming experience in other languages stems from Dr. A.J.

Perlis. He suggested a series of such introductions to various

languages. This introduction was written having in mind the FORTRAN

or BASIC programmer. The APL programmer may have some more

difficulty to understand the features of ALGOL 60 as explained

here with respect to the control statements. This introduction

to ALGOL 60 does not claim to be complete. The idea was rather

to present to the experienced programmer the flavor of ALGOL 60

and illustrate its major features. The author's preference and

taste certainly shows in the presented material in the sense that

goto statements and switches are not discussed at all.

The problems are closely related to the presentation and are

numbered accordingly. The Revised Report on ALGOL 60 gives the

programmer the exact description of the language and it is recommended

that he use it when there is any doubt about permissability or

correctness of program writing. The presentation is machine-

independent; the programmer should read the appropriate manual for

implementation details.
 -2

1. Data structures

1.1 Constant

1.2 Variable

1.3 Expression

1.4 Array

1.5 Procedure

1.1 Constant

type is expressed by representation

integer constants 21, -273

real constants 2.718, 3.14^-2, 1 Q - 3 , .1138

boolean constants true and false

string constants "any sequence of symbols"

1.2 Variable

A variable is represented by a name (official term: identifier).

A name does not reflect the type.

A variable is declared to be of certain type in a declaration.

A declaration consists of a declarator followed by a list of variable

names and is terminated by a semicolon.

Note: transitions to a new line or a new page or next card have no meaning

in ALGOL 60 as delimiting symbols.

integer i, j, k; real x, y, z;

Boolean a, b, c; integer n;
•4

A variable can get a value by means of an "assignment statement"

i := j + 1 ; x := 2 * y + z; a := true
 -3-

1.3 Expressions

1.3.1 Arithmetic expressions

unary operators - +

binary operators in order of priority f

* / T
+

a - b * c t 2

Operations of equal priority are evaluated from left to right.

Two arithmetic operators should not be written in succession.

a * - b is incorrect.

Parentheses overrule the priority:

a * (b + c) a * (-b).

Note: * / have equal priority.

Hence, 24/6 * 4 evaluates to 16 (and not to 1 ) .

The type of an arithmetic expression depends usually on the types of its

operands.

p+q, p-q, p*q a r e

°f tyP e
integer

if both operands are of type integer.

and of type real

if at least one of the operands

is of type real.

p / q, p't q are of type real if the operands

are of type real or integer.

p -r q is only defined for two operands

(integer division)

of type integer; the result is also

of type integer.
 -4-

When a real expression is assigned to an integer variable, its value

is rounded to the nearest integer i : = x -f 0 . 7 where x = 2.6

yields i = 3 .

1.3-.2 Relations

Relational operators < ^ = ^ are binary operators all of the same

priority, which is lower than that of + and -•

The operands are of type integer or real, the result is of type Boolean.

i > k + l x ;> 0 n < i t ( j + 2*k)

N.B. x < y £ z is illegal, because the type of x < y is Boolean and

^ requires 2 operands of type integer or real.

1.3.3 Boolean expressions

operators in priority order

—i not (unary operator)

A and

V or
(binary operators)
-> imp

= eqv

na b A - ] C V x < y + z

The type of a Boolean expression is Boolean.

The value of a Boolean expression is either true or false.

table

a b —J a aAb aVb a -> b a = b

t • t f t t t t
t f f f t f f
f t t f t t f
f f t f f t t

The value of a Boolean expression can be assigned to a Boolean variable

a: = b A c = x > 0
 -5-

1.4 Array

An array is an ordered set of variables.

The array is represented by a name.

The elements are represented as subscripted variables*

v [4] M [ 3,7] v [i + k] M [ i + 1, j - 2 ] .

The subscript expression must be of type integer or real.

Its value is rounded to the nearest integer value.

All subscripted variables of an array are of the same type. This

type is not reflected in the representation and therefore an appropriate

array declaration must be given.

integer array v[6:24];

real array M[3 101, 7:9];

:

boolean array p, q, r [0:7];

A subscripted variable can be used in places in the program where a

variable could be used.

v[i + 3] * ( v[i + 1] f2 - v[i + 2]) + 2 * i * v[i]

p[3] A a = q[6] V v p[i] : = a A y ^ z + l

v[i + j] := 2 * k + 1

1.5 Procedure

A procedure is a piece of program represented by a name. Execution

of a procedure may require some parameters. These are given in the

so called actual parameter list.

x: =SIN(0.75); y : = LN (2.718); READ(x); PRINT(i);

i: =MAX(i,j); SWAP (x,y)

Some procedures are considered to be available to any ALGOL 60

program. These procedures are said to be in the ALGOL library.

 -6-

Library procedures are

ABS LN and usually

ENTIER EXP READ
SIN SIGN WRITE
COS SQRT PRINT
ARCTAN although their use varies
from one implementation
to another.

A programmer can also write his own procedures (like MAX and SWAP).

In such cases a procedure declaration is required. This is discussed

in section 4.
 -7-

2. Control statements

Declarations are followed by one or more "statements". Statements

are executed in sequence. Statements are separated from each other

by a semicolon (since new line, next card and new page have no

meaning).

i: = 3; v[i] :=5; i := i + 1; v[i] := v[i -1] + 2 *

v[i - 1] t2

2.1 Repetition statements

for i:= exprl step expr2 until expr3 do S

i must be a variable of type real or integer (usually the latter).

(Don't forget that i must occur in a declaration!)•

exprl, expr2, expr3 must be expressions of type real or integer.

(a constant or a variable is also a form of expression).

for i := 50 step 1 until 90 do v[i] : = 0;

The elements 50 through 90 of vector v are set to zero.

for i := 0 step 1 until n do

for j := 0 step 1 until i - 1 do a[i,j] := a[j,i]

s:=l; for i := 1 step s until 20 do s + 1

The statement s := s + 1 is executed for i = 1, 3, 6, 10, 15 successively.

The final value of s = 6.

n : = - 14;

for i :== 1 step - 2 until n - 1 do n := n + 2

The statement n := n + 2 is executed for i = 1, -1, -3, -5, -7

successively. The final value of n = -4.

2.2 Compound statement

Any sequence of statements can be grouped together into one statement

 -8-

by surrounding the sequence with the bracket pair

begin end

for i 3 step 1 until 10 do

begin k := v[i] ;

v[i] = v[18 - i ]
: ;

v[18 - i] = k
:

end

2.3 Conditional repetition

for i := expr while BE do S

i must be a variable of type integer or real.

expr must be an expression of type real or integer.

BE must be an expression of type Boolean.

Statement S is executed repeatedly as long as BE is found true. The

value of expr is assigned to variable i every time BE is evaluated.

i := 12;

for i = i + 1 while i £ 20 do
: v[i] = SQRT (v[i])
s

A very useful conditional repetition statement has been implemented

in more recent ALGOL 60 systems.

while BE do S

BE must be an expression of type Boolean.

If this statement is not available in the implementation you work

with (it is not in ALGOL 60 officially), it can be programmed using

the for while statement with a dummy assignment.

for a : = a while BE do S

Let a and b be positive integers. The greatest common divisor of a

and b is computed by
 while r / 0 do (or for a : = a while r / 0 do)

begin

q:= a -r b;
r:= a - q * b;
a:= b;
b:=r

end;

gcd:= a

Conditional statement

if BE then S

BE must be an expression of type Boolean.

Statement S is only executed if BE is true.

Restriction: S may not be itself a conditional statement. This

rules out the sequence ... then if ...

This restriction is not a serious one, because any statement can be

transformed into a compound statement by surrounding it with begin end,

if i < 20 then begin if v[i] < 0 then v[i] := -v[i] end

for i := 12 step 1 until 36 do

if v[i] $ 0 then begin PRINT(i); PRINT (v[i]) end

This statement prints the index and value of the elements unequal zero,

Selective conditional statement

if BE then Si else S2

BE must be an expression of type Boolean. Either statement Si or

statement S2 is executed depending on whether BE is true or false.

Restriction: Statement Si may not be a conditional statement, nor

a repetition statement. Statement S2 may be any kind of statement.

 -10-

As a thumb rule one had better avoid always the sequences ... then if

and ••• then for •••

The restrictions are not serious, since the problem can be avoided by

using a compound statement,

i^f n > 5 then

begin for i : = 5 step 1 until n do

if v[i] = i then v[i] : = 0

end

Suppose it is known that array v has positive elements numbered from

0 through 512, The value of the largest element is computed by

max := 0;

for i : = 0 step 1 until 512 do

if max < v£i] then max : = v[i]

 -11-

3. Blockstructure

A block has the form

begin < declarations > ; < statements > end

The names declared in this block are said to be "locals . of this

11

block.

A block is a statement.

A whole ALGOL program has the form of a block.

Declarations are terminated by a semicolon.

A statement is only terminated by a semicolon if it is not followed

by else or end.

A block differs from a compound statement in that it contains

declarations. When a block is entered, the declared names are

added to the ones declared in surrounding blocks. When the end

of a block is passed, the names declared in this block are deleted.

Thus, the names exist only as long as the block is executed.

begin integer i, j; real array A, B, C [0: 10, 0: 10];

for i : = 0 step 1 until 10 do

for j : = 0 step 1 until 10 do READ (A[i,j]);

for i : == 0 step 1 until 10 do

for j := 0 step 1 until 10 do READ (B[i,j]);

for i : = 0 step 1 until 10 do

for j : = 0 step 1 until 10 do

begin integer k; real sum;

sum : = 0;
for k := 0 step 1 until 10 do

sum := sum + A[i,k] * B[k,j];

 -12-

for i : = 0 step 1 until 10 do

for j : = 0 step 1 until 10 do PRINT (C[i,j])

end

It is allowed to program lower- and upperbound of an array declaration

in the form of arithmetic expressions with program variables.

real array A[0:n], B[n;2 * n + 1];

But the value of n must be defined when this declaration is processed.

This implies that n cannot be declared in the same block as arrays

A and B, since all declarations of a block come before any of its

statements and so,no value can have been assigned to n in this block

when the array declaration is processed. This problem is solved by

using an outerblock:

begin integer n; READ(n);

begin real array A[0:n], B[n: 2 * n + 1];

end

end

It is allowed to declare a name in an inner block that already was

declared in an outer block.

begin integer i; real x,y;

READ(k); READ(y); i : = 3;

if x > y then

begin real i;

i : = x;

x = y;
:

••e

y : = i;

end;

i := i + 1

end
 -13-

When the statement i: = i + 1 is executed, the real variable

i does not exist any more. The list of defined names is in the

outerblock innerblock

real y real i

real x real y

integer i real x

////////// integer 1

///////////

The location of a name is found by searching the namelist from the

top down for the first occurrence of that name. Thus, in the

innerblock real i is found and in the outerblock integer i.

 -14-

4. Procedures

4*1 In the matrix multiplication example the statement to read in arrays

A and B are very similar. A procedure to read in any array with

boundpair list 0:10, 0:10 is

procedure MREAD(W); real array W;

begin integer i, j;

for i : = 0 step 1 until 10 do

for j : = 0 step 1 until 10 do READ (w[i,j])

end;

The first line is the so called procedure heading. It consists of

the declarator procedure, the procedure name, the (list of) formal

parameter(s) and the so called specification of the parameter(s).

The procedure heading is followed by the so called procedure body

which must be a statement and which is usually a block.

Procedure MREAD can easily be modified into the more general procedure

SQMREAD which reads any square matrix:

procedure SQMREAD (W, a, z ) ; array W; integer a, z;

begin integer i, j;

for i : = a step 1 until z do

for j : = a step 1 until z do READ (W[i,j])

end;

A procedure is used in a program in the form of a procedure call.

In the matrix multiplication example the statements to read arrays

A and B could be replaced by the procedure calls:

SQMREAD (A, 0, 10);

SQMREAD (B, 0, 10)

if the declaration of procedure SQMREAD was placed in the program

following the declaration of the arrays A, B and C.

 -15-

4.2 Type procedures

Type procedures serve to compute the value of a function.

y:= SIN(1 - x ) ; c = 2 * MAX(a,b) + 1;

:

a:= MAX(MAX(x,MAX(y,z)), MAX(u.v))

If the used type procedure is not in the ALGOL library, it must be

properly declared in the program.

real procedure MAX(x,y); real x,y;

begin

if x > y then MAX:= x else MAX: = y

end;

integer procedure NORM(v,n); real array v; integer n;

begin integer i; real sum;

sum: = 0;

for i:= 1 step 1 until n do

sum:= sum + v[i]*v[[i];

N0RM:= SQRT (sum)

end;

boolean procedure PRIME (x); integer x;

begin integer i; boolean pr;

" pr:= X - 2 * ( X T 2 ) / 0 ;

i:= 1;

for i:= i + 2 while i f2 ^ x and pr do

• if x - i * (x ? i ) = 0 then pr: = false;

PRIME: = p r

end;

The procedure name of a type procedure is used in the procedure

 -16-

body as a kind of local variable to which the procedure result is

assigned. The procedure name may only be used as a local of the body

left of the assignment operator:= (or <-)• It would for instance

lead to incorrect programming when in the last procedure the boolean

variable pr were replaced by PRIME because of the expression

i 12 ^ x and pr

It is allowed that a non-local variable is used in a procedure body.

If a value is assigned to such a non-local variable the procedure is

said to have side effects. These may be very useful, but one should

realize the consequences.

begin integer i;

integer procedure P(x); integer x;

begin P:= x; x:= x + 1 end;

i:= 4; i:= P(i) * i

end

The last value assigned to i in this program is 20.

 -17*

Recursion

The rule for constructing names (identifiers) in ALGOL 60 is given in

terms of a so called recursive definition.

<identifier>: := <letter>|<identifier> <letter>|<identifier> <digit>

The symbol | is read as "or".

From this definition one can derive that an identifier is a non empty

sequence of letters and digits of which the first is a letter. The

definition is called "recursive 11

because the term identifier is

defined in terms of itself.

There is a class of functions that can be defined recursively.

For instance f(a,n) = a 11

a ^ 0, n ^ 1 is defined by f(a,l) = a and

f (a,n) = a * f(a,n-l)

A recursive function can be computed by means of a recursive procedure

i.e., a procedure that calls itself in its body.

real procedure P0WER(a,n); real a; integer n;

begin

if n == 1 then POWER: = a

else POWER := a * P0WER(a, n-1)

end;

real procedure SUM(v,n); real array v; integer n;

begin

if n = 1 then SUM: = v[l]

else SUM:= v[n] + SUM (v, n-1)

end;

The coefficients of Pascal's triangle fulfill the relations

G,n = 1 for n ^ 0
n
for n 2> 0, 1 ^ i ^ n-1
 -18-

integer procedure PCOEF(n,i); integer n,i;

begin

if i = 0 or i = n then PC0EF:=1

else if i ^ 1 and i ^ n-1

then PC0EF:= PC0EF(n-l, i-1) + PC0EF(n-l, i)

else PC0EF:= 0

end;
 -19-

Call by name and call by value

In the last example the formal parameters n and i are used several

times in the procedure body. To be more precise: the values of n

and i are used several times.

The ALGOL 60 rule for using formal parameters is: when a procedure

body is executed, the substituted actual parameter is evaluated

every time when the corresponding formal parameter is encountered.

This means that on account of the call

PCOEF (a t 2 + 1, 3)

the expression a f 2 + 1 is evaluated every time when the value of

n is needed.

This reevaluation of the actual parameter is clearly not needed in

the given example. It could easily be avoided by using local

variables in the procedure body:

integer procedure PC0EF(n,i); integer n, i;

begin integer locn, loci;

locn:= n; loci:= i;

if loci = 0 or loci = locn then COEF := 1

else if loci^l and loci ^ locn -1

then PC0EF:= PC0EF(locn -1, loci - 1) + PC0EF(locn -1, loci)

else PCOEF:=0

end;

It is, however, not necessary to program such an optimization

explicitly with local variables. The programmer has the option of

indicating in the declaration of a procedure that a particular parameter

has to be evaluated only once. It is said that such a parameter is

called by value. The way to do this is to add to the procedure heading

 -20-

a so called value list, in which the parameters are specified that

are "called by value. 11

integer procedure PCOEF (n,i); value n, i; integer n,i;

begin

if i = 0 or i f n then PCOEF:= 1

else if i £ 1 and i £ n-1

then PC0EF:=PC0EF(n-l,. i-1) + PCOEF (n-1, i)

else PCOEF: = 0

end;

Formal parameters called by value should be considered as local

variables of the body which will be initialized when the procedure

is called with the value of the corresponding actual parameters.

If a parameter is not listed in the value list, it is understood that

the corresponding actual parameter is substituted for the formal

parameter (and not just its once computed value). This use of a

parameter is refered to as "call by name".

To clarify the difference of call by name and call by value this example:

begin integer i; integer array v[l:100]} integer a,b;

integer procedure VSUM(u) ; value u; integer u;

begin

integer s; s:= 0;

for i:= 1 step 1 until 100 do s:= s + u;

VSUM:= s
 -21-

integer procedure NSUM(u); integer u;

begin

integer S; S:= 0;

for i:= 1 step 1 until 100 do S:= S + u;

NSUM:== S

end;

for i:= 1 step 1 until 100 do v[i]:= i;

i:=5;

a:= VSUM(v[i]);

b:=NSUM(v[i])

end

When VSUM is called, v[5] is evaluated and u gets (only once) the

value 5. Variable a gets the value 500. When NSUM is executed v[i]

is substituted for u and the result is that b gets the value 5050*

Epilogue.

Various concepts of ALGOL 60 and several details have been left out

of this introduction. The most important ones left out ar^e: goto

statements, switches, forlists, ngiltiple assignments. One can find

those in other descriptions of the* language as for instance in the

ALGOL.60 manual for the Univac 110* or the PDP 10.

A complete and accurate description of ALGOL 60 is found in the

Revised Report on the Programming Language ALGOL 60

P.Naur (ed).
published in Rosen's book,
The Communications of the ACM (Jan 63)

and various other places.

It should be stated that in cases where this introduction deviates

from the Report, the latter should be considered as the correct

description.
 -22-

Bibliography on ALGOL 60

Manual

[1] PDP-10 Algol Manual or UNIVAC 1108 ALGOJ

Definition

[2] Naur, P. (ed.) Revised Report on the Algorithmic

Language ALGOL 60. Comm.ACM 6
(Jan 63).

[3] Knuth, D. E. The remaining trouble spots in

ALGOL 60. Comm.ACM 10 (Oct. 67).

[4] Abrahams, P. W. A final solution to the dangling

else of ALGOL 60 and related
languages. Comm.ACM 9 (Sept 66).

[5] Knuth, D.E., Merner, J.N. ALGOL 60 Confidential. CACM, Vol. 4,

1961.

Philosophy

[6] Perlis, A. J. The synthesis of algorithmic systems.

J.ACM 14 (Jan 1967).

History-Bibliography

[7] Bemer, R. W. A politico-social history of ALGOL.

Annual Review In Automatic Program-
ming, 5 Pergamon Press, 1969.

[8] Sammet, Jean Programming Languages: History and

Fundamentals, Prentice-Hall, 1969.

Introductory

[9] Bottenbruch, H. Structure and use of ALGOL 60.

J.ACM 9 (Apr 62).

[10] Higman, B. What everybody should know about

ALGOL. Computer Journal 6 (1963)
p. 50.

[11] Ekman, T. and Froberg, C. Introduction to ALGOL programming.

Oxford University Press, (1967).

[12] Dijkstra, E. W. A Primer of ALGOL 60 Programming,

Academic Press, London, 1962.

T
 -23-

Imp lamentation

[13] Evans, A. An ALGOL 60 Compiler.

Annual Review in Automatic Program-
ming, 4. Pergamon Press (1964).

[14] Randell, B. and Russell, L.J. ALGOL 60 Implementation.

Academic Press, (1964), 418 pp.

[15] Dijkstra, E« W. '•Making A Translator for ALGOL 60,"

Annual Review of Automatic Program-
ming, Vol. 3., MacMillan, 1963,
pp. 347-356.

Extensions

[16] Wirth, N. A Generalization of ALGOL. Comm.ACM

6 (Sept 63).

[17] Perlis, A.J. and Iturriaga, R. An extension to ALGOL for manipulating

formulae. Comm.ACM 7 (Feb 64).

[18] Wirth, N. and Weber, H. EULER: A generalization of ALGOL

and its formal definition. Comm.
ACM 9 (Jan, Feb. 66).

[19] Wirth, N. and Hoare, C.A.R. A contribution ot the development

of ALGOL. Comm.ACM 9 (June 66).

[20] Dahl, O.J. and Nygaard, K. SIMULA - An ALGOL-based simulation

language. Comm.ACM 9 (Sept 66).
 -24-

1. Data structures

1.1 Cons tants

integer constants 216 1 -273 +1296

real constants 3.1415^2 2.718&-4

+10.10&-4 -12.24&f3

42.1516 -0.07236

.541778 +.0321927

.23&-2 -.032&8

&5 -&-7

Which of the following list represent numbers in ALGOL 60?

3.1&0.5 2. 252 3.5.6 3.14&(-2)

1.2 Variables

types: integer real Boolean

Examples of names: low, spring, found, easel, case2, switchlto3

Are the following names correct identifiers in ALGOL 60?:

onethrunine a6

lthru9 a(6)

bow farrow a [6]

1.3 Expressions

begin integer i,k; real a,b,c; integer

bow, arrow; real target; Boolean hit;
a := 3; b := a+1; c : = b+2.3&-l;
PRINT(a); PRINT(b); PRINT(c);
i :=* a; k := b/a*c; hit := a < b and c ^ 2;
PRINT(i); PRINT(k); PRINT(hit);
hit := hit or b < 1 £ a - 2.4/(b*a)*ct2 s hit;
PRINT(hit)
end
 -25-

The values assigned to the variables are successively

a: 3.0
b: 4.0
c: 4.23
i: 3
k: 6
hit: true true

Which values are assigned to the variables in the following assign-

ment statements?

begin
integer i,j,k; real a,b,c; Boolean p,q;
a := 3.1415&H;
b := 2.718&-1;
c := 1.01;

P := a < b and c > 10*bf2;

q : = a-b < 3*c or -i p;
i := a+b; j := a+c; k := b+c;
i := j+2*a; j := i-k*cT2; k := b*(i-l)-a;

P := i > j and j > k s i> k;

q := a/l2*bt2 < c+l/(b/(2*c+D)
end

2. Control statements

2.1 Counting repetition

(1) to read in the elements of a matrix stored rowwise on an input

file

begin integer i,j; real array A[l; 100, 1:100];

for i :=* 1 step 1 until 100 do
for j 1 step 1 until 100 do READ(A[i, j ])
 -26-

(2) to zero out the upper triangle of a matrix

begin integer i,j; real array A[l: 100, 1: 100];

for i 1 step 1 until 100 do
for j := i+1 step 1 until 100 do A[i,j] = 0
:

(3) Horner's scheme for computing the polynomial

n n-1
a x + a -x + ... -f a,x + a^
n n-1 1 0

is to start with y = a and repeat n times

y := y*x+a i for i = n-l,n-2,...,0

begin integer i; real array A[0:20]; real x,y;

for i := 0 step 1 until 20 do READ(A[i]);
READ(x);
y A[20];
for i := 19 step -1 until 0 do y := y*x+A[i]

(4) Write a piece of program that computes the sum of the elements

on the main diagonal of a given square matrix of 144 elements.

(5) Write a piece of program that assigns to a boolean variable "none 11

the value true or false according to whether none of the elements

of a given vector v of 25 components is zero or not.

(6) Write a piece of program that computes the inner product of two

given vectors u and v of 80 components each.

(7) Write a piece of program that computes the length of a given

vector of 45 components.
 -27-

(8) Write a piece of program that gets as input a positive integer

less than 1,000,000 and prints successively the digits of the

octal representation of this number.

(9) The coefficients of an expansion of (l+x) are given by

n

C[n,0] = 1 for n £ 0, C[n,k] = £ * C[n-l,k-l] for 1 £ k £ n.

Write a program that computes C[24,13].

2.2 Compound statements

(1) To transpose a square matrix A of 169 elements

for i := 1 step 1 until 13 do

for j := 1 step 1 until i-1 do
begin
x := A[i,j];
A[i,j] := A[j,i];
A[j,i] := x
end

(2) The sequence of Fibonacci is defined by

F(0) = 0, F(l) = 1 and F(n) - F(n-l) + F(n-2) for n £ 2

to compute the 25th Fibonacci number

begin integer a, b, f, i; a := 0; b : 3
1;
for i := 2 step 1 until 25 do
begin f = a+b;
b = a;
a = f
end

end
 -28-

(3) Matrix A represents 15 vectors of 20 components each; to compute

a matrix D that contains the distances between all pairs of

vectors in A.

begin integer i,,i,k; real s; real array A[1:15, 1:20], D[l:15,l:15];

for i :- 1 step 1 until 15 do
for j := 1 step 1 until 20 do READ(A[i,j]);
for i :- 1 step 1 until 15 do
begin D[i,i] := 0;
for j := i+1 step 1 until 15 do
begin s := 0;
for k := 1 step 1 until 20 do
s := s + (A[i,k] - A[j,k]) t 2;
s := SQRT(s); D[i,j] := s; D[j,i] := s
end
end
end

(4) Matrix A represents 45 vectors of 12 components each. Write a

piece of program that assigns to the Boolean variable "sphere"

the value true or false according to whether all 45 vectors are

equal in length or not.

(5) Write a piece of program that multiplies matrices A[0:24,0:9]

and B[0:9,0:11] and stores the result into matrix C of proper size.

(6) A number sequence is given by s[0] = 0, s[n] = n+s[n-l] for n ^ 1.

Write a program that computes s[36].

 -29- 4

3 Conditional repetition

(1) A root a of the equation f(x) » x can be found by means of suc-

cessive approximation if in the neighborhood of x 3

a

ABS(f (x)) £ k < 1.

f
The approximation starts with a value x^

in that neighborhood and a new approximation is computed from

the last according to =* f ( x

n ^) • 0 n e c a n
prove using the

first mean value theorem that the sequence x^ converges to a.

F(x) = 0.01(x 4
+ 2x 3
- 3x 2
- 4)

F(x) intersects the x-axis between x=l and x=2

begin real a, newx, x; x := 1; newx := 1.5;

while ABS(newx-x) > 0.0005 do
begin x := newx;
newx := ((x+2)*x-3)*xt2*0.01-0.04
end
end

(2) Queues q ^ ^ * * • • ><lg ^ e stored in array A[l:1000]«

a
The first

element of q ± is A[i]. If A[k] » n, it means that A[n] is the

next element of the queue to which A[k] belongs. The last

element of a queue has the value zero, unused elements of A have

the value -1.

To delete the last element of queue q^ (assuming the proper

declaration and initializations)

k := i; next := i;
while A[next] > 0 do begin k : = next; next := A[k] end;
A[k] :« 0; A[next] :=* -1
 -30-

(3) To assign the value true to Boolean variable "zero" if and only

if vector v[1:2048] has at least one element equal zero, and,

if so, the index of the first element = 0.

i := 0; zero := false;
while i < 2048 and —i zero do
begin i := i+1; zero := v[i] - 0 end

(4) Write a program that gets a real number x as input and computes

n terms of the expansion 1-x + xt2/2J -xt3/31+... until a term

with absolute value ^ 0.0005 is found.

(5) Newton s method to approximate a solution a

1 fl M
of f(x) = 0 is

based on the computation of the sequence x ,x-,x ,... where x

n

U 1 z o

is a point in the neighborhood of point a and

x^ ^ - x
+ - f(x^)/f (x^).
1
Write a program to approximate

6tl/3 with an accuracy of 0.001 applying Newton's method on

f(x) » x - 6 .
3

(6) Write a program that gets as input a positive integer n less

than 1000 and assigns to the Boolean variable "prime 11

the value

true or false depending on whether or not n is a prime number.

(7) Write a program that computes the largest Fibonacci number less

than 8000.

(8) Write a program that joins all the queues of the example

presented earlier into one queue q^ leaving out the first elements

^2'^3 * * * *'^9*
 -31-

(9) Integral (f(x),a,b) is computed by successive approximation.

The first approximation is

i(0) (0.5*f(a) + 0.5*f(b))*(b-a)

Approximation i(k+l) is obtained from i(k) by taking the mid-

points of every sub interval and adding the approximation of

all intervals. So,

i(l) := 0.5*(f(a) + f(m))*(b-a)/2 + 0.5*(f(m) + f(b))*(b-a)/2

i(l) := i(0)/2 + f (m)*(b-a)/2 where m = (a+b)/2

or
i(2) := 0.5*f(a) + f(am))*(b-a)/4 +

0.5*<£(am + f (m) )*(b-a)/4 +

0.5*(f(m) + f(mb))*(b-a)/4 +

0.5*(f(mb) + f (b))*(b-a)/4

or
i2 := i(l)/2 + (f(am) + f (mb))*(b-a)/4
where am is the midpoint of segment (a,m) and bm of segment (m,b).
Write a program that computes Integral(x*sin(x),0,1) until two

approximations differ less than 0.001

2.4 Conditional statement

(1) Vector v[l:100] has positive elements. To compute the value of the

largest element

begin integer i; real max; real array v[l:100];

max := 0;
for i 1 step 1 until 100 do
if max < v[i] then max := v[i]

end
 -32-

(2) a[l:50] and b[l:50] are sets of integer numbers. The sets are

called "disjunct" if for all i,j - (1,...,50) either a[i] < b[j]

or a[i] > b[j]. To assign to Boolean variable "disjunct" the

value true if and only if sets a and b are disjunct. (Assume

proper declarations and initialization.)

mina : = a[l]; maxa : = a[l]; minb := b[l]; maxb := b[l];

for i := 2 step 1 until 50 do
begin if mina > a[i] then mina := a[i];
if minb > b[i] then minb := b[i];
if maxa < a[i] then maxa := a[i];
if maxb < b[i] then maxb b[i]
end;
disjunct := maxa < minb or njaxb < mina

(3) v is an ordered set of 512 numbers. To find out up until which

element the values are in ascending order (assume proper declara-

tions and initialization)

i := 1; ascending := true;
while i £ 512 and ascending do
begin i i+1;
if v[i] < v[i-l]
then begin i := i-1; ascending false end
end

(4) v is a set of 100 "names". A "name" is represented as a positive

integer less than 100,000. Write a piece of program to arrange

the-names in alphabetical order, i.e., in ascending order.

(5) v is a set of 729 integer numbers. Write a piece of program

that computes the minimum value of these numbers and how many

elements of v are equal to that minimum.

 -33-

(6) v[l:120] is a set of real numbers. The "diameter" of a set

is MAX(v[i] - v[j]) for all i,j - (1,...,120). Write a piece

of program that computes the diameter of v.

(7) Matrix A represents a set of 15 vectors of 6 elements each and

matrix B represents a set of 12 vectors of 6 elements each. The

"distance" between A and B is the minimum of all the distances

between a vector of A and one of B. Write a piece of program

that computes the distance between sets A and B.

2.5 Selective conditional statement

(1) The maximum of four variables a, b, c and d can be computed in

one statement

if a > b
then begin if a > c
then begin if a > d then max :» a
else max := d

end
else if. c > d then max := c
else max := d

end
else if b > c
then begin if b > d then max := b
else max := d

end
else if c > d then max : = c
else max := d
 -34-

(although the following sequence is simpler:

max := a;

if max < b then max := b;

if max < c then max :- c;
if max < d then max := d)

(2) A group of 75 students got scores on a progress test with a scale

ranging from 1 to 100. The results are regrouped into four cate-

gories: excellent, good, average, insufficient. The boundaries

are at 90, 70 and 50. To program how many of each category there

are:

begin integer excellent, good, average, insufficient;

integer i, score;
for i := 1 step 1 until 75 do
begin
READ(score);
if score ^ 90 then excellent := excellent + 1
else if score ^ 70 then good :=* good + 1
else if score ^ 5 0 then average :- average + 1
else insufficient := insufficient + 1
end;
PRINT(excellent); PRINT(good); PRINT(average); PRINT(insuffici
end

(3) Does the following program compute correctly the greatest common

divisor for any pair of integer numbers (a,b)?

T
 -35-

begin integer a,b,gcd;

READ(a); READ(b);
while b / 0 do
if a > b then a := a-b
else b := b-a;
gcd := a; PRINT(gcd)
end

(4) Let f(x) be a continuous function on interval [a,b] and let f(x)

intersect the x-axis exactly once in that interval. The point

of intersection is approximated by shrinking the interval to

l/2, l/4, l/8, ... of its size successively, preserving only

that part in which the zero point lies. Write a piece of pro-
4 3 2

gram that solves the equation 4x + 3x - 2x - 1 = 0 with this

method.

(5) Array A[l:25] is considered as a circular storage area for a

queue q (i.e., each element A[i] has a successor; the successor

of A[25] is A[l]). Elements are added at the front of q and

elements are deleted from the rear of q. Write pieces of program

to add to q an element just read in and to delete an element

from q, assuming that the variables "front" and "rear" point to

the correct elements of A #

(6) Someone wrote in an ALGOL 60 program

 -36-

begin integer i,j; integer array a[l:100]; Boolean stop;

for i := 1 step 1 until 100 do READ (a[i]);
stop := false; i := 0;
while —i stop do
if i £ 100 and a[i] > 0
then begin PRINT(a[i]); i := i+1 end
else stop true
end

His program worked all right if he supplied it an input file

with at least one 0 or negative number, but it gave a runtime

error for an input file with only positive numbers. A friend

suggested to replace the ^f statement by the following statement:

if i ^ 100
then begin if a[i] > 0
then begin PRINT(a[i]); i := i+1 end
else stop := true
end

else stop := true

and this cured the problem! Try to explain why.

Block structure

(1) The first number on an input file represents the number of components

of the vectors and is followed by the components of two vectors.

To write a program that computes the inner product

 -37-

begin integer n;
READ(n);
begin real array u,v[l:n]; integer i; real inprod;
for i := 1 step 1 until n do READ(u[i]);
for i := 1 step 1 until n do
begin
READ(v[i]);
inprod := inprod + v[i]*u[i]
end;

end
end

(2) To program the linear transformation v := A*v where v is a

vector of n components and A is the n*n transformation matrix.

Blockstructure is used to limit the scope of intermediate

variables.

begin integer n; READ(n);

begin real array v[l:n], A[l:n, l:n];
begin real array u[l:n];
begin integer i;
for i 1 step 1 until n do READ (u[i]);
end;
begin integer i,j >

for i 1 step 1 until n do

for j := 1 step 1 until n do READ(A[i,j])
end;
begin integer i;
for i 1 step 1 until n do
begin integer j; real sum;
sura := 0;
for j := 1 step 1 until n do sum := sum + A[i, j ]*u[j ];
v[i] := sum
end
end
end
end
end
 -38-

To write a program that solves a set of linear inhomogeneous

equations of which the matrix is triangular with a non-zero

main diagonal, i.e.,

to solve A*x = e

where a,, = 0 for all 1 £ i < j £ n and

a. . 4 0 for all 1 £ i £ n.
11 '

Since it is triangular, it pays to store the matrix into a one

dimensional array B[l:n*(n+1) r 2] to avoid a waste of storage.

The general mapping rule is A[i,j] = B[i*(i-1) + 2+j ]

begin integer n; READ(n);

begin real array B[l:n*(n+1) + 2],x,C[l:n];
begin integer i, upb; upb := n*(n+l) -r2;
for i := 1 step 1 until upb do READ(B[i]);
for i := 1 step 1 until n do READ(C[i])
end;
begin integer i;
for i :=» 1 step 1 until n do
begin
x[i] := C[iJ / B[i*(i+1) + 2 ] ;
begin integer k;
for k := i+1 step 1 until n do
begin integer m; m := k*(k-l) + 2+i
C[kJ C[k] - B[m] * [ i ] ;
x

B[m] :« 0
end
end
end
end
and
end
 -39-

(4) Write a program that computes the coefficients of the product

of two polynomials. The input file contains the index of the

first polynomial followed by its coefficients, followed by the

index of the second polynomial and its coefficients.

(5) C is a vector of n elements. Matrix A is a n*n matrix of which

only the elements a[i-l,i] for 2 ^ i £ n, a[i,i] for 1 £ i £ n

and a[i,i+l] for 1 £ i ^ n-1 are not equal to zero. Write a

program to solve the set of linear equations

A*x = c

The input file contains successively n, the 3n-r2 non-zero

elements of A and. the elements of c.

(6) Array A[0:10,000] is used to store variable size segments of

at least two array elements. The first contains the link,

i.e., a pointer to the segment next in order of ascending array

address (zero for the last segment); the second contains the

segment length in number of array elements. A[0] points to the

first segment. Segments are added and deleted and so, after a

while, unused areas are scattered through array A. Write a

program that moves the segments to one end of array A (and up-

dates the links properly) such that only one free area results

at the other end of array A.

 -40-

(7) The Boolean product and Boolean sum of two integers represented

as binary numbers is computed by operating on corresponding bits

of those numbers according to the following tables:

Boolean product Boolean sum

0 * 0 = 0 0 + 0= 0

0*1 = 0 0 + 1=1

1 * 0 = 0 1 + 0=1

1 * 1 = 1 1 + 1 = 0

Thus, 101101 * 011001 = 001001

101101 + 011001 = 110100

The Boolean multiplication is coded on the input tape as 1 and

the Boolean addition as 0. Write a program that gets as input

the operation code followed by two (decimal) integer numbers

£ 1,000,000, that converts the numbers into binary representa-

tion, carries out the indicated operation and prints the sequence

of resulting bits.

(8) Write a program that prints out in how many ways, and how, one

can pay an amount of n cents (n ^ 100,000) in dollars, quarters,

dimes, nickels and cents.

(9) Write a program that prints the triples (a,b,c) of Pythagoras

(i.e., the relation between a, b and c is at2 + bt2 = ct2),

for 1 £ a < b £ 100.

 -41-

Procedures

(1) To write a Boolean procedure that gets the value true if and

only if a given square matrix is orthonormal.

Boolean procedure ORTH(A,n); value n; real array A; integer n;

begin integer i,j; Boolean ok;
real procedure INPROD(rowi, rowj, n ) ;
value rowi, rowj^ n; integer rowi, rowj, n;
begin integer i; real sum; sum := 0;
for i 1 step 1 until n do
sum sum + A[rowi, i] * A[rowj,i];
INPROD := sum
end;
ok := true; i := 0;
for i : = i +1 while i £ n and ok do
ok := INPROD(i,i,n) = 1;
i := 0;
if ok then
begin for i := i+1 while i £ n and ok do
begin j i;
for j j+1 while j £ n and ok do
ok := INPROD(i,j,n) = 0
end
end;
ORTH := ok
end;

(2) A push-down stack

lf 11
is a list of elements to which elements are

added and from which elements are deleted at one end of the list

called the stack top. Let array A[0:1000] be used as storage

area for a stack and let integer variable "top 11

point to the top
 -42-

of the stack. A stack element consists of one or more elements

of array A, the topmost element contains the length in number

of array elements. To write a procedure that adds a stack ele-

ment stored in v[l:n] to the stack.

procedure PUSH(v,n); value n; real array v; integer n;

if top + n ^ 1000 then WRITE( stack overflow")
!l

else begin integer i;

for i := 1 step 1 until n do A[top+i] := v[i];
top := top-frrt-1;
A[top] := n
end;

To write a procedure that computes et(-x) with an accuracy of

0.0005 and counts the number of terms needed to achieve this

accuracy

real procedure EP0W(x,n); value x; real x; integer n;

begin real term, sum; integer sign;
term := 1; sum := 1; sign := 1; n := 1;
while ABS(term) > 0.005 do
begin sign := -sign;
term := sign*term*x/n;
sum := sum + term;
n := n+1
end;
EPOW := sum
end

One can also pass a function (procedure) name as parameter.

To write a procedure that uses the method of bisection to ap-

proximate the zero point of a function in a given interval with

a given accuracy
 -43-

real procedure ZERO(f,a,b.acc); value acc;

real a,b,acc; real procedure f;
begin real m;
if f(a) > 0 then begin m := a; a := b; b := m end;
while b-a > acc do
begin m :=(b-a)/2;
if f(m) < 0 then a := m else b := m
end;
ZERO := m
end;

Suppose the program in which procedure ZERO is declared contains

also the declaration

real procedure F(x); value x; real x;

begin F := (2*x+l) * xt2 - 2 end;

The program could use the procedures for instance in the statement

y := ZERO(F,0,1,0.0005) or

y := ZERO(SIN,-.5,.5,0.00001)

The sequence of Fibonacci numbers is defined by

F(0) = 0, F(l) - 1, F(n) - F(n-l) + F(n-2) for n £ 2.

Write a recursive procedure FIB(n) which computes F(n).

Write a procedure F(exp,x,n) that tabulates the values of an

expression represented by exp for a set of n arguments represented

by x.
 -44-

(7) Write procedure POP which deletes an element from the stack

(see problem 2 ) .

(8) What is the result of the following program?:

begin integer i,j,k;

procedure P(F); procedure F;
begin integer a;
procedure Q; begin i := i+1 end;
j j+i;
if j = 1 then P(Q);

a := i+k;
PRINT(a)
end;
procedure S; begin k := k+1 end;
i := 0; j := 0; k := 0;
P(S)
end

(9) Palindromes. A palindrome is a vector V of values such that

V = XY where X - reversal of Y. E.g., 110011. Write a Boolean

function that determines if a vector is a palindrome. Write

another which determines if a vector consists of a list of

palindromes; e.g., 110110.

(10) How well do you understand call-by-name and call-by-value?

 -45-

begin real A,B;

real procedure INCV(X); value X;real X;
begin XHC+1; INCVHC end;
real procedure INCN(X);real X;
begin XHC+1; INCNHC end;
real procedure ADDV(Y);value Y;real Y;
ADDV-Y+Y;
real procedure ADDN(Y);real Y;
ADDN«-Y+Y;

A H ; B-ADDV(INCV(A));
comment A IS NOW , B IS NOW

A H ; B«-ADDV(INCN(A));
comment A IS NOW , B IS NOW

AH.; B«-ADDN(INCV(A));
comment A IS NOW

A H ; B«-ADDN(INCN(A) ) ;
comment A IS NOW , B IS NOW
end;

Miscellaneous

(1) Sieve of Eratosthenes

Eratosthenes method to list the prime numbers less than 1000

1

is the following one:

Set up the sequence 1,2,3,..•,1000.

• Scratch out 1.

Repeat:
Scratch out all multiples of the first number not yet
scratched out (for the first time this number is 2) until
to the right of this number all numbers are scratched out.
 r

-46-

Write a program that lists the prime numbers less than 1000 in

ascending order using the sieve of Eratosthenes.

(2) Continued fractions

let q x = 1/(1+1)

q 2 = 1/(1 + 1/(1+1)) etc.

as i -* infinity, q^ q = 0.6183 ...

Write a procedure PHI(n) that will return the value q^.

Write two versions of PHI, one recursive and one iterative.

(3) The QUEEN's problem

To place on the chessboard 8 queens in such a way that no.queen

can capture another queen in one move. Write a program that pro-

duces all solutions of this problem (there are 92 solutions).

(4) Write a procedure SWAP(x,y) that exchanges the values of x and

y. Check that it works for SWAP(i,a[i])and SWAP(a[i],i).

(5) 32 locations numbered from 0 through 31 are arranged in a circle

(i.e., location 0 follows location 31). The locations are to

be filled with either 0 or 1 in such a way that all numbers from

0 through 31 are represented exactly once by interpreting all

possible contiguous sets of 5 locations as the representation

of a binary number. Write a program that produces all solutions

of this problem.

t
 -47- i

(6) Tower of Hanoi

Given three pins and a pack of n disks on the first pin. The

disks are stacked in order of descending diameter. Move the

disks to pin 2 (possibly using pin 3 as intermediate storage)

so that (1) the disks end up in the same order on pin 2 as they

were on pin 1, (2) at no time a larger disk is on top of a smal-

ler one and (3) only one disk at a time is moved. Write a pro-

gram that solves this problem for an arbitrary number of disks.

(7) (J. Weizenbaum)

Write a program that, for given positive n, determines the small-

est number z that can be decomposed into the sum of two n-th

powers in at least two non-trivial ways.

For instance: n = 3 z - 1729 = lf3 + 12|3

= 9t3 + 10t3

n = 4 z = 635318657 = 59t4 + 158t4

- 13314 + 134 T4