The BCS Professional Examination

Certificate
April 2005
Examiners Report
Software Development
General
This paper frequently involves developing code logically from first principles. While memorised
code may afford speed in writing some answers, in others there is invariably a difference between
the memorised terms and those of the question, which leads to difficulty. It also leads to inclusion
of additional material not asked in the question. The examiners do not penalise this directly as
they consider the time wasted in writing it out as a sufficient penalty.
It is very useful that candidates are given practice in answering questions on paper. Developing
programs interactively at a PC or terminal is a poor preparation for this examination. While such
practical work is useful in developing students understanding of a programming language it is NOT
good for developing algorithms, or developing a part of a much larger program.
Flowcharts are inferior to pseudo-code for developing algorithms. They consume much time in
drawing, imply GO TO instructions and are difficult to develop logically. They become unwieldy
for a problem of significant size or complexity.
No marks are given for copying out the question, nor for repeating declarations given in part a,
say, in the latter part of the question. Algorithmic development need only show those statements
or sections which are developed further, provided that it is clear where they belong, rather than
copying all of it out again.
Question 1
1.

The file stock contains details of books currently held in stock by a retailing bookshop:

<ISBN>
10-digit integer
0 7493 97543

<Author>
20 characters
Bernieres, L.

<title>
40 characters
Capt. Corellis mandolin

<stock level>
4-digit integer
25

<price>
real number
7.99

Another file supply contains details of books held in a warehouse:

<supplier name>
20 characters
Vintage Fiction
a)

<ISBN>
10-digit integer
0 7493 97543

<available>
1 char, T/F
T

<stock level>
4-digit integer
650

<price>
real number
6.55

Specify suitable data structures to hold details for both of these files. Hence write a suitable file
description of stock and supply.

(6 marks)

b)

The <Author> and <title> fields both use many characters and for a typical file with 1000 books would result
in waste of storage space, as much of the available space would not be used. Suggest how this storage space
could be minimised without losing processing capability.
(4 marks)

c)

Write an initial algorithm to read the stock file once and for each entry search the supply file to match the
ISBN of a book. If the book is found in the supply file and the price lower than on the shop stock file,
print the full details of the book. Count how many records are on the stock file, how many matches are
found and print the results at the end.
An initial algorithm and at least one stage of development must be shown.

1 (of 17)

(20 marks)

Answer Pointers
a)
Possible data structures are
TYPE stocktype = RECORD
isbn : longint
author : array [1..20] of CHAR;
title : array [1..40] of CHAR;
stocklevel : shortint;
price : REAL;
END;
TYPE supplytype = RECORD
isbn : longint;
supplier : array [1..20] of CHAR;
available : BOOLEAN;
stocklevel : shortint;
price : REAL;
END;
b)
There are several possibilities. Some procedural languages (like FORTRAN) allow dynamic
declaration of arrays, so the actual size is not fixed and only storage space needed is used. A
linked list could be used here in PASCAL so that much space is saved where there are short titles
and the occasional very long title wont cause an overflow. It would also be possible to use
temporary files.
c)
Initial algorithm:
assign files
set counters to zero
WHILE NOT end-of-file on stock file
READ stockfile record
increment stock record count
WHILE NOT end-of-file on supplier file
BEGIN
READ suppler record
test for match of isbn from both files:
IF match THEN BEGIN
increment match-count
IF {available AND lower price} THEN PRINT record details
END
END
Final counter output
END.

2 (of 17)

Development:
ASSIGN stock, supply files, open for reading {always system dependent}
rec_ct 0
match_ct 0
WHILE NOT EOF(stock) DO
BEGIN
READ stock, stock_item
rec_ct rec_ct + 1
match FALSE
WHILE NOT EOF(supply) AND (NOT match) DO
BEGIN
READ(supply, supply_item)
IF stock.isbn = supply.isbn THEN
BEGIN
match_ct match_ct + 1
match TRUE
IF {item is available} AND supply.price < stock.price
THEN {PRINT book details}
END
END {inner while loop }
RESET supply-file
END {outer while loop }
WRITE number of records on stock file, rec_ct
WRITE number of matches, match_ct
CLOSE supply file, stock file
END
Examiners Comments
Marks were given for correct structure (however shown) as well as appropriate instructions. No
development at all was penalised, as this was specifically required by the question.
Students should be able to choose the correct type of loop for a particular problem. Thus a FOR
or count-controlled loop is inappropriate here. The initial algorithm should have been checked
before developing the code and any errors or omissions corrected.
Excessive use of simple procedures containing one or two instructions are to be avoided.
A wide range of ability was shown in part c). Mostly students had some idea of the correct
program structure; marks were given for correct components with more for correctly-terminated
loops. Only a few wasted time by repeating declarations made in part (a). About half of the
answers showed little or no development of the initial algorithm, for which added-value marks
were awarded. If there was only one algorithm in a developed (even coded) stage marks were
given for every valid instruction and loop.

3 (of 17)

Question 2
2.

line
number
1
2
3
4
5
6
7

Program best-of-three
main variables a, b, c : INTEGER
Procedure select (t1,t2,t3)
BEGIN
If t2 > t1 THEN swap (t2, t1)
If t3 > t2 THEN swap (t3, t2)
If t2 > t1 THEN swap (t2, t1)
END
INPUT (a, b, c)
select (a, b, c)
WRITE (a, b, c)
END

a)

Dry run the algorithm with input values 1 3 2 at the top level. Use the line numbers in your answer.
(12 marks)

b)

Certain languages (e.g. Pascal) do not incorporate a system procedure swap. Write appropriate code for
this procedure and state where it would be placed in the supplied code, with reasons.
(7 marks)

c)

What property must the variables used for procedure parameters (t1, t2, t3) have for the desired result to be
achieved?
(4 marks)

d)

In some languages the variables (a, b, c, t1, t2, t3) can be strings of characters. Where could this be applied?
What constraints are needed to make this work properly?
(7 marks)

Answer Pointers

a)
line number /instr
5 / Input
6/ procedure call
1 / procedure code
2 / IF t2 > t1
3 / IF t3 > t2
4 / IF t2 > t1
End procedure code
6 / procedure return
7/ Write
end

test
?
?
?
true
true
false

t1
?
?
1
3
3
3

t2
?
?
3
1
2
2

t3
?
?
2
2
1
1

output

3 2 1

b)
Procedure swap (x,y)
local temp : INTEGER
BEGIN
temp x
xy
y temp
END
Procedure swap would be placed above the procedure select code in those languages which
require procedures to be placed before the code that invokes them. Thus procedure select is
placed before the top level which invokes it.

4 (of 17)

c)
The variables must be capable of passing back the changed values held in them after the code
has been run as well as initially receiving initial values from the top level variables. Hence invoke :
a t1, b t2, c t3 and on return : t1 a,
t2 b, t3 c.

d)
This could be used in word-processing code to exchange characters in written documents.
Problems would be the matching up of the strings of characters, particularly if these are of variable
length. A constraint as to the maximum length of such characters would have to be found, such
that space was not wasted by having the same length as the largest string for all the character
variables. Much more code would be needed especially if blocks of such characters are swapped.

Question 3
3.

a)

Write a procedure, in Pseudocode, or Structured English or a programming language with which you are
familiar, that implements insertion in an array of data ordered in an ascending manner as follows:
The procedure should accept an integer parameter EL that is the data to be inserted.
The procedure should scan an array (name = LIST, size = 100).
The first element in the list contains the number of data elements it contains. That is, LIST[1] = number of
elements in LIST.
All existing data in LIST is held together, that is all data is stored adjacent, ascending.
On finding where to insert the new data, you will need to make a space and move the existing data to create
an empty space.
If LIST is full, that is, there is no space to insert the new value, an error message should be given. If
successful, no return message need be given.
(20 marks)

b)

Dry-run your pseudocode from a) above with the following data:

EL = 23
LIST =
10

4
1

15
2

22
3

24
4

98

99

100
(10 marks)

Answer Pointers
a)
The procedure should accept an integer parameter EL that is the data to be inserted.
The procedure should scan an array (name = LIST, size = 100).
The first element in the list contains the number of data elements it contains. That is,
LIST[1] = number of elements in LIST.
All existing data in LIST is held together, that is all data is stored adjacent, ascending.
On finding where to insert the new data, you will need to make a space and move the
existing data to create an empty space.
If LIST is full, so that there is no space to insert, an error message should be given. If
successful, no return message need be given.

5 (of 17)

Notes: The array has subscripts 1 to 100, not 0 to 100, or 0 to 99.

The maximum number of ascending data numbers that can be stored is 99.
When requiring the index of the last used position in the array it is LIST[1]+1.
Procedure INSERT should
Accept EL as parameter, and prepare to insert.
Check list not full, LIST[1]=99. If so, return error message.
If list is empty, LIST[1]=0, then LIST[2]:= EL, increment LIST[1] and exit.
Scan for insertion point. Scan with index, say i, from 2 until LIST[1], stopping when
EL<LIST[i] or we exhaust the scan range. The insertion position is now i.
Shift elements LIST[i+1] to LIST[(LIST[1]+1)] up one place. A loop, say with index j set to
run down from (LIST[1]+1) to i, moving each LIST[j] into LIST[j+1], and decrementing j
until j=i, the insertion point.
Now insert EL, LIST[i]:=EL
Increment LIST[1]
b) Dry run should show something like:
EL
23

LIST[1]
4

i
2 until 5
3
4
5. Stops
because
EL<24.
i is insertion
point.

j
6

LIST[i]
10
15
22
24

5
until
5
5

22

LIST[j]

j runs
(4+1)
down to
5
24
23

LIST[j+1]

24

5
5
1

10
2

15
3

22
4

23
5

24
6

98

99

100

Examiners Comments
The simple approach given in the solution for part (a) is sufficient. There are more complicated
algorithms that would be acceptable but they would cost candidates valuable time to write down.
For example finding the correct position by binary chop or adding the new number at the end and
then coding a complete sort of the array.
There is a nice, simple algorithm that starts at the end of the array (position 5 in the example) and
works backwards moving elements one place up until the place for the new element is found and
then the new element is just added in the last created 'hole'.
The "list full" test was often written LIST[1] equals 100. In fact the LIST is full when LIST[1] equals
99.

6 (of 17)

Many answers did not work if the list was empty initially so there should have been special code to
test for an empty list and insert EL as the new (only) data element.
The question asks for a procedure with parameters. The procedure is to work on a LIST which
already contains some data. Thus no answer should have had any input or output, reading or
writing.
The data in LIST was given and was already sorted. Thus no answer should have started by
sorting the LIST.
No answer should have had any pointers or records.
The names LIST, EL were given in the question so no answer should have called these items by
any other name.
The code to move a sequence of elements one place further up the list must start with the highest
numbered subscript and work down to the lowest. If the move is done from the lowest subscript
upwards, the result is that the 'moved' elements will all be the same value (the first one 'moved')

Question 4
4.

a)

Algorithmic design in Structured Programming uses three principles. Describe EACH of these principles.
(12 marks)

b)

Give an example of each principle applied to CODE.

(9 marks)

c)

Give an example of each principle applied to DATA.

(9 marks)

Answer Pointers
a) Structured programming uses principles
Sequence,
Selection and
Iteration.
b)

Sequence is code is begin..end block, or similar in other programming languages.

Selection in code is if..then..else.. , or similar in other programming languages.
Iteration in code is a loop.

c)

Sequence in DATA is a record or structure, or similar in other programming languages.

Selection in data is a variant record, or union or alternate structure, or similar in other
programming languages.
Iteration in DATA is an array or file.
Many candidates answered parts (a) and (b) and either repeated their answer to part (b) in
part (c) or left part (c) blank so 9 marks were lost immediately. Data structure diagrams are a
convenient way of showing sequencing, selection and repetition in data

7 (of 17)

+---+
|seq|
+---+
/ \
+---+ +---+
|
| |
|
|
| |
|
+---+ +---+

+---+
|sel|
+---+
/ \
+---+ +---+
| o| | o|
|
| |
|
+---+ +---+

+---+
|rep|
+---+
|
+---+
| *|
|
|
+---+

Examiners Comments
Candidates seemed able to describe selection and repetition but were unable to find any words to
describe sequencing. However it was not sufficient to say that "selection means selecting..." or
that "repetition means repeating...". When describing a term T you have to find words different to
T.
In section (b) many candidates gave very elaborate examples, often complete programs. For
selection something simple like "if x>y then x:=y" or "if(x>y)x=y;" would have been enough.
In many answers the examples requested in part (b) were actually written in part (a). Strictly
speaking the marks for the examples should not have been awarded unless they were written in
the correct section, but in this case the examiner was lenient.
Some unsatisfactory answers gave a definition of the Software Development Life Cycle or a
description of an algorithm.

Section B
Question 5
5.

Briefly describe the type of testing you would use for:

a) Unit tests
b) Integration testing

(6 marks)
(6 marks)

Give reasons for your answers.

a) For unit tests, use logic tests (white box testing where the code is available). Test cases are
designed to force toe algorithm to take all the coded paths. Sometimes a series of tests are
necessary. Output for each test must be checked against known or calculated results.
3 marks for white box testing with appropriate definition and 3 marks for further material, particularly
about how it is done with reasons, e.g. this type of testing verifies that individual units function
correctly and are appropriately coded.
b) Integration is testing the assembled units, and checking that interfaces generate and accept
correct data with fail messages being produced for incorrect data. this is usually termed black box
testing, where the code is not available. Black box tests may necessitate further white box tests of
units.
3 marks for black box or functional testing; 3 marks for description about how it is done with
reasons.

8 (of 17)

Examiners Comments
A bookwork question answered by nearly all candidates. Generally done quite well with some
good answers. Others confused black and white box testing, or used both forms everywhere.
Others wrote everything they knew about testing from memory and hoped the examiner would pick
out what was relevant. Some candidates expressed their answer in bullet-point form, which is to
be commended. Repetition of points does not gain extra marks. Also students should avoid using
the word integration in the definition of integration testing. (Integration testing is where the
modules are integrated and tested.)
Some students wrote quite brief answers and made few points beyond a basic definition like unit
testing is where the modules are tested separately. Students need practice in answering this kind
of question, and to realise that more than that is expected for 6 marks!
Students must realise that at this level unselective regurgitation of memorised notes will attract
fewer marks than answers which supply relevant material. Simple logic would help here and avoid
errors in recall like Unit testing needs black box testing and integration testing needs white box (or
is it the other way round)? Thus integration testing is for the whole system (parts put together)
hence impractical to do white box testing which looks at logic and actual code - there would be a
lot to do! Hence use black box testing, which looks at interfacing and input/output.

Question 6
6.

An inter-island ferry has a passenger-carrying capacity of C people. In general W people are waiting for the ferry.
If W > C then those left behind are P percent of those waiting and L percent of the ferrys capacity.
a)

Derive separate expressions for P and L in terms of the variables (C, W). Simplify these expressions as far
as possible.
(6 marks)

b)

Write a program
i)
to accept a value for C
ii) to accept lower and upper limits for W
iii) to accept a value for W and
iv) then print a table with the headings below

People waiting
(W)

Number of people left

behind

% of people waiting left

behind (P)

% of ferrys (L) capacity

left behind

(6 marks)

9 (of 17)

Answer Pointers
a)
W C = P - fraction of people waiting left behind (if C > W this is zero)
W
W - C = L - fraction of people left to capacity of ferry.
C
P = 100 ( W C ) = 100 ( 1 - C / W )
W
L = 100 ( W - C ) = 100 ( W / C - 1 )
C
Substituting for C we get P = 100(

} these equations express P and L in terms of C and W

L )
100 + L

Substituting for W we get L = 100 (W / C 1)

b)
ferry program
100
110
120
130
140
150
160
170
180
185
210
215
220
230
240
250
260
300
350

PRINT "ferry passenger capacity program"

PRINT "Implemented in Qbasic"
PRINT "input ship passenger capacity/legal safety limit"
INPUT C
PRINT "input LOWER limit for number of people waiting for ferry"
INPUT low
PRINT "input UPPER limit"
INPUT high
PRINT "total waiting left behind percent
capacity percent"
REM 'FOR' loop to print table between input limits
FOR W = low TO high STEP 5
IF C > W THEN LET B = 0 ELSE LET B = W - C
IF C > W THEN LET L = 100 * W / C ELSE LET L = 100 * (W / C - 1)
LET P = 100 * L / (100 + L)
PRINT TAB(5); W; TAB(17); B; TAB(30);
PRINT USING "###.#"; P; TAB(45);
PRINT USING "###.##"; L;
NEXT W
END

Examiners Comments
a) required explicit equations for P and L in terms of input values W and C. Some candidates did not
realise this and had a page of algebra; others managed it quite easily.
b) wanted a table of values for W and corresponding P, L values. Thus a loop between input
maximum and minimum values for W was needed with (W -C), P and L values calculated from the
equations derived in (a) was necessary. Incorrect equations derived in (a) were not penalised in (b).
To get all the marks proper captions such as input the number of people waiting for the ferry rather
than input W were expected. Again answers ranged from completely correct to the bizarre.

10 (of 17)

Question 7
7.

The diagram below represents a linked list with a single pointer field.
99
Head

80

member-1

70

member-2

70

member-3

end-member

a)

Write a declaration for one member of this list in a suitable language. State which language you are using.
(3 marks)

b)

Copy the diagram and show the necessary pointer movements to exchange member-2 with member-3.
(5 marks)

c)

Write code for these pointer movements in the same language you used in part a).

(4 marks)

Answer Pointers
99

head

member-1

80

70

70

member-2

member-3

end-member

The diagram represents a linked list with a single pointer field.

a)
TYPE

ptr = ^ node;
node = RECORD
data : INTEGER;
next : ptr
END;

PASCAL used

b)
2
member-1

member-2

member-3

1
3
c)
temp := member1.next
member1.next := member2.next
member2.next := member3.next
member3.next := temp

{address needed in stage [3]}

{stage [1] }
{stage [2]}
{stage [3]}

Examiners Comments
It is essential that the pointer movements are made in the correct order, otherwise the address in
one field altered too early is wiped out before it is needed in a later phase.
To exchange these members within a linked list it is obviously necessary to traverse the list until
member-1, member-2 and member-3 are in the right places.
11 (of 17)

Question 8
8.

False

True
a>b

x>y

True

FFalse

Statement-1
Statement-3

Statement-2

a)

Represent the flowchart (above) in pseudocode.

b)

Represent the pseudocode below by a flowchart.

IF test-1 THEN

IF test-2

THEN statement-1
ELSE statement-2

ELSE statement-3
c)

(4 marks)

(4 marks)

Specify appropriate test data for b) so that each of the three statements is output. State which output should
be produced for each set of test data.
(4 marks)

12 (of 17)

Answer Pointers
a)

flowchart
False

True
test-1

True
statement-1

test-2
False
statement-3

statement-2

a) IF test-1 THEN statement -1

ELSE IF test-2 THEN statement-3
ELSE statement-3

13 (of 17)

b)

test-1
test-2
statement-3
statement-2
statement-1

c)
Only the simple truth table given below is wanted here.
test-1
true
true
false

test-2
true
false
(true OR false)

output/execution
statement-1
statement-2
statement-3

Question 9
9.

a)

Describe one purpose of an index-access file? Give reasons for your answer.

b)

When in operation, is it better to keep the index in main memory, or on disk? Give reasons for your answer.
(6 marks)

14 (of 17)

(6 marks)

Answer Pointers
a) An index on a file is for speed of access. Its purpose is to respond to random requests for
records such as from a telephone query.
b)

Keeping the index in memory makes for faster access because electronic speed are higher
than disk-based electro-mechanical speeds.
Keeping index on disk retains integrity in event of system crash.
Either conclusion acceptable so long as it is supported.

Examiners Comments
Many answers for part (a) did not score good marks because the answer was a description of an
index-access file rather than a purpose of having an index-access file. So detail about how the
index is created, how it is updated, etc did not score marks because it did not tell the purpose of
having an index.
In both parts of this question many candidates used the word "easier" in a way that was not
accepted. Using an index on a file is not "easier". It is harder. There is more work to do in making
the index and keeping it up to date. Accessing a record using an index is not "easier". The
"easiest" thing to do is start at the beginning and look through all the records until you find the one
that you want. The benefit that you get with an index is speed, you can find records "faster".
Candidates need to be careful with terminology. There is a big difference between "looking for a
file" and "looking for a record".
When giving examples it is not enough to say something like "An example of an index-access file
is a telephone bill". A batch production of telephone bills could work well on a sequential file. The
example that requires index-access is responding to a telephone enquiry which requires random
access to the customer database in order to be answered quickly.
Something about the way part (b) of this question was worded confused many candidates. The
question in the sentence is not "When ... ?" but "is it ... ?"
Note that a diagram like
+---+---+---+---+
| a | c | d | j |
+---+---+---+---+
is not acceptable as a diagram of an index. It is necessary to have key/address pairs.

Question 10
10.

Show a GUI design for a system that is used to monitor 4 different locations using four remote cameras. Include
in your design GUI elements of a drop-down list and a set of radio buttons, and give them appropriate functions to
perform. Explain your design fully.
(12 marks)

Answer Pointers
Expect a screen with four windows. The drop-down list should be for some function that is itself a
rather long sequence. This might be a degree of camera movement (up-down, left-right), or
similar. The radio-button should be for some data that represents a choice from alternatives. This
might be brightness, or a zoom ratio.

15 (of 17)

Examiners Comments
This question asked for a DESIGN, so there were no marks for an answer which simply described
the purpose of the various interface elements. The question asked for the design to be
EXPLAINED fully, so a diagram on its own could not obtain full marks.
It was not acceptable to have the drop-down list and the radio buttons used for the same purpose
e.g. both allowing the selection of one camera out of 4.
It was clear that some candidates were visualising 16 cameras with 4 in each location.
In a drawing/diagram only one entry in a drop-down list is visible so it is vital that the other entries
are shown or described in the explanation.
Among the answers there were designs which involved choosing a camera on one page, then
choosing a zoom on another page and then a brightness on another page, etc. These were not
considered good designs as it seemed better to have all the control situated on a single page.

Question 11
11.

Describe the functions of the following software elements. State whether each is system software or application
software.
a) A Compiler
(4 marks)
b) A Scheduler
(4 marks)
c) A Word Processor
(4 marks)

Answer Pointers
a) Compiler is system, and translates source to executable.
b) A scheduler is system, and sequences tasks for execution on the CPU.
c) A Word Processor is Application, and helps users to format text and diagrams in a document.
Examiners Comments
It was not enough to say that a compiler compiles things, that a scheduler schedules things, a
word processor does word processing, etc.
For part (b) some candidates gave answers relating to "cron jobs" or similar and some candidates
gave answers relating to "diary" type programs.
Several answers focussed on relatively unimportant, secondary features of the software
concerned. It is not correct to claim that the principle task of a compiler is to find syntax errors in a
program. It is not correct to claim that the principle task of a word processor is to make line
drawings.

Question 12
12.

Briefly describe the type of documentation you would assemble for the maintenance of a desktop database
application. Give reasons for your answer.
(12 marks)

Answer Pointers
For maintenance, need to document the internal design, workings and tests for the functions so
that changes can be made.
Also, the documentation should include design relating to the database (ERD or SQL), Test
specifications, and Test results.

16 (of 17)

Also, the documentation should include details of how the application was generated - version
number and date.
Examiners Comments
This question was quite poorly answered in general.
It was not sufficient just to describe documentation in general. It was not acceptable to answer the
question "Why produce documentation?". It was not acceptable to answer the question "How is a
software product documented?".
The question asked about a specific kind of application. Full marks could not be obtained unless
the answer was written specifically for MAINTENANCE and specifically for a DESKTOP
DATABASE application.

17 (of 17)