Commentary - Free resources for teaching

chairwomanlettersΛογισμικό & κατασκευή λογ/κού

13 Νοε 2013 (πριν από 4 χρόνια και 1 μήνα)

210 εμφανίσεις

GCE 2003
June Series
Report on the Examination
Advanced Subsidiary – (5511)
Advanced – (6511)
Computing
￿ Advanced Subsidiary
￿ Advanced
Further copies of this Report on the Examination are available from:
Publications Department, Aldon House, 39, Heald Grove, Rusholme, Manchester, M14 4NA
Tel: 0161 953 1170
or
download from the AQA website: www.aqa.org.uk
© Assessment and Qualifications Alliance 2003
COPYRIGHT
AQA retains the copyright on all its publications. However, registered centres for AQA are permitted to copy
material from this booklet for their own internal use, with the following important exception: AQA cannot give
permission to centres to photocopy any material that is acknowledged to a third party even for internal use
within the centre.
Set and published by the Assessment and Qualifications Alliance.
The Assessment and Qualifications Alliance (AQA) is a company limited by guarantee, registered in England and Wales
3644723 and a registered Charity 1073334. Registered address Addleshaw Goddard, Sovereign House, PO Box 8,
Sovereign Street, Leeds LS1 1HQ.
Kathleen Tattersall, Director General.
CONTENTS
Computing
AS Units
Page No.
CPT1 Computer Systems, Programming and Networking
Concepts..............................................................................................4
CPT2 Principles of Hardware, Software and Applications..............7
CPT3 Practical Systems Development................................................12
A2 Units
Page No.
CPT4 Processing and Programming Techniques..............................17
CPT5 Advanced Systems Development..............................................20
CPT6 Coursework......................................................................................24
Mark Ranges and Award of Grades ......................................................................29
Report on the Examination Advanced Subsidiary & Advanced - Com
p
utin
g
￿￿￿
4
Computing
CPT 1: Computer Systems, Programming and Networking
Concepts
Candidates need to be made aware that proprietary names of hardware and software never score
marks. Acronyms are only accepted in exceptional circumstances. Candidates should provide full
names for the terms they are using.
Question 1
Candidates find it very difficult to express themselves clearly and concisely to define computing
terms.
System software is the layer of software which enables users to operate the computer hardware, such
as an operating system. Many candidates wrongly think that software already loaded on the system at
time of purchase is system software.
Application software is any program written to perform an end user task, such as word processing
software.
Special purpose application software is software written to solve a particular problem, such as
accounts software.
Bespoke software is software specially written for an organisation with the advantage of the software
exactly matching the user's requirements and the disadvantages that the organisation will have to wait
for it to be written and it will not be as tried and tested as readily available software. Many candidates
wrongly state that bespoke software is more expensive to write. It may be more expensive to the client
because the cost of production can not be spread over a wider user base.
Question 2
A significant number of candidates do not know that 1024 bytes make 1 Kilobyte. Even fewer
candidates could state correctly that the largest pure binary integer that can be stored in 2 bytes is
65,535 (or 1111 1111 1111 1111 1111). Incorrect responses ranged from as low as 3 to many
thousands.
The bit patterns asked for were largely well answered, but candidates should be made aware that
leading zeros are required when bit patterns to a specified length are asked for. The whole purpose of
binary is that only two states, 0 and 1, can exist.
The concept of parity checking eludes many candidates. Few could explain how a computer system
would use a parity bit. The parity bit is set when the character is first generated, by adjusting the
parity bit to make the number of 1s even (for even parity). Then the parity is checked at the receiving
end and if the parity is now odd an error has occurred. The question clearly stated that in the given
computer system the parity bit was the most significant bit of each byte. However, many candidates
only looked at the parity across the whole 16 bits. Many candidates were not able to translate the
characters 3 and 7 into ASCII codes with the code ranges for digits explicitly provided by the
question.
Report on the Examination Advanced Subsidiary & Advanced - Com
p
utin
g
￿￿￿
5
Question 3
Many candidates scored well on this question. However, a significant number did not take account of
the arrow heads into and out of the system, and VDU controller and keyboard controller labels were
interchanged.
The stored program concept generally is very poorly understood. The machine code instructions
(program) and the data are stored in main memory and instructions are fetched sequentially and
executed by the processor. Programs in main memory can be replaced by other programs at any time.
Many candidates wrongly stated that this concept means that programs were stored in Read Only
Memory or in secondary storage. It seems that many candidates are not clear about the difference
between secondary storage and main memory, even when they had correctly labelled the diagram
earlier in the question.
Question 4
Some candidates provided excellent answers while others seemed to not have the most fundamental
understanding, suggesting that sounds are turned into either a 0 or a 1. Analogue sound is converted
into digital by sampling the height of the analogue wave at regular intervals and representing the
measured value by a binary code. Some candidates wrongly suggest that a modem can be used for
this. To convert the digitally recorded sound so it can be played back through speakers, most
candidates suggested a 'sound card' which was accepted, though what was required was the essential
component of the sound card, which is the Digital to Analogue Converter. MIDI was a frequent
wrong answer.
Vector graphics have featured in pervious examination questions in this unit but few candidates could
give a full answer. The (x,y) co-ordinates of the two endpoints of the line are stored, together with
information about its thickness, colour and what type of shape it is. Many candidates could not
express clearly enough the idea of start and end points and others referred to pixels, clearly not
understanding the different methods of representing images.
Question 5
In part (a), many candidates failed to gain both marks available because they did not explicitly state
that nodes must be connected and that a Wide Area Network extends over more than one site.
In part (b) most candidates correctly stated that a modem is required to connect a stand-alone
computer to the Internet, but some wrongly suggested a network card. Most candidates correctly
identified a browser as the application software required to access a web site, though candidates did
not gain credit if they only stated a proprietary name such as Internet Explorer.
In part (c) an explanation of leased line networking and dial-up networking has been asked many
times before, but still many candidates concentrate on the method for payment of such services rather
than on the difference of connection, even though many providers currently charge a flat rate for both
services. Leased line networking provides a permanent connection, while with dial-up networking the
users have to re-connect every time they want to go on-line.
For part (d), serial transmission was still poorly described. Candidates should refrain from mentioning
bits of data as this may well jeopardise credit in the future. A creditworthy definition is: ‘Bits are sent
one after another along a single wire’.
Part (e) showed that baud rate is a very poorly understood concept. Even though most candidates have
come across the term in the context of Internet communication, very few seemed to know what it
exactly measures. One baud is the number of signal changes per second. Bit rate is the number of bits
Computing - Advanced Subsidiary & Advanced Re
p
ort on the Examination
￿￿￿
6
transmitted per second. As one signal may encode more than one bit, bit rate is not necessarily the
same as baud rate.
Question 6
Candidates must be aware that one word answers are not usually sufficient. Just spelling out the full
word from the abbreviation in the URL is enough to explain what the different parts can tell us.
Particularly vague were the answers to the history part of the URL, which is a folder name of the web
site. Some candidates wrongly thought it was a link. Some candidates suggested, wrongly, that each
part of the URL is coded into a set of digits and this is how the IP address is arrived at. Each web site
has a unique IP address. The more user-friendly domain name is resolved to its IP address through a
Domain Name Server. A creditworthy response was that each domain name maps onto one IP
address. (Although it should be noted that several domain names may map onto the same IP address.)
Very few candidates correctly stated the range of possible IP addresses. Many did not appear to see
the significance of each group of digits being stored in one byte and therefore the possible range was
0.0.0.0 – 255.255.255.255. Some candidates were on the right track but failed to see that the largest
possible integer that can be stored in a byte is 255 and stated a variety of other values. This is an
important concept to understand the pending change to longer IP addresses as the world runs out of
the currently available addresses.
Question 7
In part (a), a large number of candidates failed to differentiate correctly between a selection statement
and iteration.
For (b), very few candidates correctly identified the given subroutines as 2 functions and one
procedure. Even though the description of the subroutines in the question stem should have made
candidates realise that both copy and concat were the same type of subroutine, many seemed to hedge
their bets and opted for one of each type. Many other candidates got the choice exactly the wrong way
round.
In (c), the response to the dry run was much more promising, suggesting that candidates are getting
more opportunity to practice this type of skill. A large number of candidates were able to follow the
algorithm with only the S2 value causing any problems.
Report on the Examination Advanced Subsidiary & Advanced - Com
p
utin
g
￿￿￿
7
CPT 2: Principles of Hardware, Software and Applications
General
Candidates with breadth and depth of understanding performed well on this paper. The same
candidates communicated this understanding effectively. They did not give one word responses but
gave fully qualified answers that revealed their depth of knowledge and understanding. For example,
in question 8(g)(ii), the better candidates answered that when a mobile phone serial number is
transmitted or enters the system it is possible that its digits can be corrupted. The check digit enables
this corruption to be detected. The weaker candidates answered that it ensures that the mobile phone
number is correct or they confused a check digit with something else altogether. The weaker
candidates often revealed poor memorisation of standard definitions. For example, the definition for
batch processing offered by weaker candidates could have been interpreted as describing an
interactive processing system or a real time processing system. A batch processing system has several
characteristics. Processing does not begin until all the data have been entered. Processing then
proceeds from beginning to end without human interaction (although human intervention is possible if
a printer runs out of paper). Although memorisation of this definition is recommended in its parts the
definition embodies the understanding of what is meant by batch processing. It is this understanding
that it is important for a student to acquire.
It was clear from question 9(b) that many candidates had not acquired knowledge of a third generation
programming language, or, if they had, they were not able to use it. Data types given referenced data
types found in a database system. For example, RandomNumber field was given the data types
Number or Long Integer and ArrivalTime the data type field Date/Time. These correspond to data
types found in Microsoft Access. Experience of a third generation programming language would
teach a candidate that data type identifiers do not use a space or spaces or the forward oblique stroke.
Third generation programming languages also differentiate between whole numbers and numbers with
a fractional part by using a different data type for each.
The file processing question (4(d)) was answered much better on this occasion with most candidates
gaining between three and six marks. By providing a template candidates responded with pseudo code
and not with prose, as was the case in previous examinations on this topic. Most candidates borrowed
pseudo code from the example in part 4(c) which was the intention. The pseudo code given in
question 4(c) was deliberately simplified. The expected response for the purpose of the update process
was to copy the old user ids file records to a new user ids file unless the next transaction file record’s
user id matched the next old user ids file record’s user id. This would occur under certain
circumstances. Put another way, the purpose was to delete old user id records under certain
circumstances. In addition, the old user ids file was taken off line after the update process. Many
candidates, in their answers, inferred a purpose that could not be concluded from the given pseudo
code. For example, many answers stated that this update process changed user’s passwords.
However, the same candidates were able to score at least three marks in part (d). The simplified
pseudo code did seem to make the better candidates think very carefully and there was evidence on
scripts of hand tracing of the pseudo code. However, this seemed to produce a beneficial outcome on
candidates’ responses to part (d), as was intended.
Many candidates were unable to clearly explain the meaning of direct access and serial file
organisation to the satisfaction of the examiners. File pipes/data streams and how data is organised for
access in file systems are very important to an understanding of computing in a range of contexts. The
content of the AS specification provides a foundation for future study in this area. For example,
backing up to a removable medium such as a magnetic cartridge uses serial file organisation.
Transmission across any serial data link such as the Internet relies upon end links that have to do
some serial processing of data streams. Database systems have not made file-based systems
redundant.
Computing - Advanced Subsidiary & Advanced Re
p
ort on the Examination
￿￿￿
8
Many candidates were unable to distinguish a fact from a benefit. In question 3(a)(ii), especially,
many candidates simply stated a fact about the scenario described in the question stem. For example,
candidates answered that the information gained about a user’s musical interests could be used to
target advertising at the user. This is a fact not a benefit. The benefit to the software company is that
the costs of advertising (marketing) could be reduced if the marketing is targeted or revenue from
sales could be increased.
Very few candidates were able to offer an acceptable advantage for a database over a separate,
independent file approach in question 1(b). Surprisingly, some candidates also experienced difficulty
producing a creditworthy validation control used in a database system.
Overall an impression was gained that, for some centres, candidates had a poor knowledge and
understanding of the content of the specification. These same candidates had not made use of previous
question papers and their mark schemes in their preparation for this examination. Candidates had not
learnt the content of the specification thoroughly. Good, old-fashioned learning of definitions by heart
had not been practised.
Question 1
For part (a), the popular answers were data type, presence, length and range check. The commonest
unacceptable answer was input mask. An input mask shows how a data value should be entered into a
field. An input mask is more applicable to the form field through which the data item is entered.
However it does not validate, for example, that a date is a valid date. A date would normally be stored
in a date field and the validation would be a data type validation. Also input masks can contain
characters that are not part of the data item. For example, the underscore character is used to indicate
a position to be filled and brackets are used in a telephone number which are not part of the data item.
Format or picture checks were acceptable, however.
For part (b), candidates struggled to give an advantage which was appropriate or accurately to state
one that was. The most popular answer was easier to search for information. Several candidates
answered that there would be no duplication when the appropriate response was that there would be
less duplication.
Question 2
Part (a) was answered well by the majority of the candidature. In some cases candidates lost a mark
because of their inability to express themselves clearly in English. These candidates gave answers that
simply echoed the stem of the question and therefore added little that conveyed understanding. A
local disk drive is one attached to the desktop computer often located inside its case. A networked
disk drive is a disk drive on another computer that is accessible to this desktop computer by virtue of
both being connected to the same network.
Part (b) gave many candidates the opportunity to score full marks, i.e. eight out of eight. However, it
was clear that some candidates had limited understanding of directories, directory structures and little
or no memorable experience of a command line interface. “Type C:” was considered by some
candidates to describe the logical drive in the question. This is incorrect. The term logical has the
meaning as perceived by a user. An operating system hides the complexities of the hardware from a
user and enables a user to refer to on-line storage area by a letter followed by a colon, e.g. C:. The
word “Type” represents an operating system command.
Report on the Examination Advanced Subsidiary & Advanced - Com
p
utin
g
￿￿￿
9
Many candidates answered part (c) correctly. Those that did not often lost the mark because
they added an oblique stroke to the directory name entered into each blank box.
Question 3
Candidates were able to identify a benefit to the user such as additional information not recorded on
CD cover may be available from the software company or information related to user interests could
be supplied. However, for a benefit to the software company, many candidates failed to supply
enough to gain a mark. Targeted marketing was often identified without relating the marketing to
saving of costs or the generation of extra revenue.
Part (b) was well answered with many candidates quoting “invasion of privacy”.
Question 4
The majority of candidates answered correctly that the file size is 30000 bytes for part (a).
Many candidates answered part (b)(i) successfully with quicker to send a two-byte integer across
network, less storage space required for a two-byte integer, easier to verify that correct password
supplied and more secure in transit and when stored being popular answers.
Fewer candidates scored full marks for part (b)(ii). The better candidates conveyed the idea that the
alphanumeric string’s characters needed to be converted into an equivalent numeric form, e.g. ASCII
codes. Next, these numeric codes needed to be combined in some arithmetic way. Finally, the numeric
result needed to be mapped onto a two-byte integer, e.g. by using integer remainder division.
Many candidates were able to explain sufficiently well to gain credit in part (b)(iii) why the hashing
function for the passwords is a one-way function.
For this examination series, given the experience of previous question papers in which many
candidates were unable to list processing steps adequately, many answering in prose instead, an
example was provided of file processing steps expressed in pseudo code. The result was a success.
Many candidates were able to use the steps given in Figure 3 to gain at least three marks in their
answer to part (d). Unfortunately, by simplifying the steps given in Figure 3, few candidates gave a
mark worthy answer for part (c). The answers given suggested that many candidates have difficulty
understanding file processing – not that candidates were confused by the simplification of the
processing steps. The better candidates were challenged by the processing steps but in doing so were
helped to produce an answer to part (d) that gained full marks.
Question 5
Candidates’ responses fell into one of two categories. If the candidates had encountered the range of
legislation they were able to score three or four marks, otherwise they scored one or zero. Most
candidates were able to answer correctly, Data Protection Act, in part (a). In part (b), the examiners
were specifically looking for a reference to Patents legislation or Intellectual Property legislation.
Some candidates quoted an abbreviated version of the Copyright, Designs and Patents Act, omitting
the reference to Patents and therefore failing to gain credit. Many candidates correctly answered
Computer Misuse Act in part (c). Other candidates misquoted the Act and referenced a Data Misuse
Act, an answer that gained no credit. Many candidates correctly identified Copyright legislation for
part (d), the better candidates giving its full name Copyright, Designs and Patents Act. Although
abbreviated answers were credited on this occasion, it is poor examination technique to state simply
Copyright in response to a question which asks for the name of the legislation. The Copyright Act is
an improvement, albeit an inaccurate one. Candidates are invited to show what they do know and
single word answers create an impression that a candidate cannot be bothered.
Computing - Advanced Subsidiary & Advanced Re
p
ort on the Examination
￿￿￿
10
Question 6
Many candidates were able to suggest a suitable method at the exit barrier for submitting the number
assigned to the ticket to the computer system. Fewer were careful enough to describe how the ticket
was assigned to the ticket at the entrance barrier. The examiners were expecting a printer to be
referenced or the action of printing in the case of a barcode, OMR, OCR, MICR and plain text
solution for writing a number to the ticket; the action of encoding or writing to the ticket in the case of
a magnetic stripe or smart card and the action of punching in the case of Kimball or Kimball-type
tags. The better candidates offered such descriptions.
A lack of experience of using a third generation programming language was exposed by part (b).
Many candidates gave data types that were lifted straight from Microsoft ACCESS and therefore
gained no credit. The question explicitly requested data types that would be available in a third
generation programming language. Candidates must understand that their study of this subject must
extend beyond products that have been intentionally designed to enable people with no background in
Computing, and no desire to be educated in this subject, to achieve practical results in the minimum of
time. For very simple practical tasks, this is desirable but as a foundation for more extensive tasks this
is a recipe for disaster. Assembling flat-pack furniture is not considered adequate enough to qualify as
a cabinetmaker.
Some candidates had difficulty selecting relevant fields and in some cases when they did choose
relevant ones lost a mark for poor choice of identifier. For example, several candidates chose Date
instead of CurrentDate. The clue was in the question stem which stated that “the computer system
remembers the Current Date, Arrival Time and Randomly generated Number.“ A principle of
software engineering is that identifiers should be meaningful and reflect the real world entities which
they represent.
For part (c)(i), the examiners were looking for the answer any record can be accessed independently
of any other record or records may be accessed in any order or a record may be accessed without
having to read any other records first. As in previous examinations, weaker candidates substituted the
word file where the word record was required thus demonstrating a lack of understanding of the
difference between a file and a record and undermining their chances of absorbing the intricacies of
file processing.
In part (c)(ii), serial file organisation is correctly described as a file in which records may be accessed
only in the order in which they were written to the file.
Many candidates for their answer to part (c)(iii) successfully answered 1000 for the number of records
that the direct access file should store at any one time. These candidates successfully concluded that a
record must exist for each parking space for direct access file organisation to work for one thousand
car parking spaces.
Fewer candidates were able to correctly answer that the random ticket number could be used directly
as the position number of the corresponding record or by an implied simple mapping used as the disk
address of the corresponding record.
Question 7
Many candidates for part (a) were able to identify two reasons why a spreadsheet is particularly useful
as a decision making tool. Support for use of formulae and graphs were the most popular answers. It
is the ability to recalculate automatically rather than to calculate automatically that is its strength.
Several candidates emphasised the latter and therefore failed to gain credit.
Report on the Examination Advanced Subsidiary & Advanced - Com
p
utin
g
￿￿￿
11
Many candidates successfully answered part (b)(i). The format of the formula was as follows:
IF($B4<$D$7,”F”,IF($B4<$D$8,”P”,”M”))
For part (b)(ii), the correct responses were Cell C4=F and Cell C5=M.
Many candidates for part (c) successfully answered that the displayed letter in cell C4 will change to
P. Consistency was required: candidates correctly answered F and M for part (b)(ii), and therefore “P”
was rejected unless candidates qualified their answer by stating without the quotes.
Question 8
Very few candidates correctly stated that a smart card is a card that is embedded with a
microprocessor and a memory chip.
Part (b) was well answered by the majority of the candidature.
Despite references in previous reports to the expected definition for a foreign key, candidates continue
to define it as “a key in one table which is a primary key in another table”. This begs the question
what is a key when it is not a primary key? The correct response is a foreign key is an attribute or
field or column in one table which is a primary key field/attribute/column in another table.
The foreign key in the CallRegister table is SimCardNo. The majority of the candidature answered
correctly.
Some candidates struggled to define accurately real time and batch processing often giving
definitions that fell short of what is expected. Updated in real time means that changes are made in a
timely manner. Batch processing is characterised by processing that only begins when all of the data
have been entered. When the processing begins it continues from beginning to end without human
interaction. However, many candidates were able to gain credit.
Part (d) (iii) asked candidates to apply their understanding of real time processing to the given
scenario. Many candidates correctly reasoned that users change location and therefore this implied
that to maintain an up-to-date record of their current location the LocationRegister must be updated
in real time.
Part (e) was answered well with credit worthy responses being to speed up searching/access/queries
and SimCardNo.
The correct answer for part (f)(i) was two, which many candidates were able to give. The QBE was
completed correctly by many candidates with only weaker candidates failing to give a credit worthy
response.
Part (g) was either known or not known. However, when known, several candidates struggled to
describe a check digit adequately. The examiners were expecting an answer that stated that it was a
digit calculated from the other digits and appended to these digits. The emphasis was on the word
calculated. Several candidates simply stated it was a digit added to the end of the
MobilePhoneSerialNo or it was a digit that checked the other digits. Both descriptions simply
repeated what was given in the stem of the question.
The purpose of the check digit is to check for/detect corruption of the other digits or the check digit
itself. The check digit alone cannot check that the MobilePhoneSerialNo is correct or valid. Neither
can it ensure that the MobilePhoneSerialNo is not corrupted. These explanations, offered by several
candidates, gained no credit. The better candidates answered both parts correctly.
Computing - Advanced Subsidiary & Advanced Re
p
ort on the Examination
￿￿￿
12
CPT 3: Practical Systems Development
General
Candidates who had created for themselves a solution which they understood, who had
included coding (annotated where appropriate) of significant steps and actual hard copies of
the reports and lists specified in the brief, did very well in this paper.
Marks were lost this year through candidates not following the brief or not reading the questions
carefully. In order to test as much of the solution as possible, different types of evidence was asked
for within one question. For example, 1 (a) (i) asked for evidence that data items had been defined or
set up; 1 a ii needed evidence that the extra data item had been used. Again in question 2 (a), coding
was required for part i, but a hard copy of the list produced for part ii. In fact, the terms ‘hard copy’
and ‘coding’ seemed not to be understood by many candidates.
Centres should appreciate that they may have grown used to a student’s handwriting but an examiner
does not have the time to do this. If a script is illegible, examiners cannot give the benefit of the
doubt to something they cannot read.
Documentation
The Principal Examiner marked scripts from one centre where every candidate had attached
documentation of more than 100 pages in length, including much that was unnecessary. It is
expected that teachers should exercise a filter of unwanted pages; over 100 is simply unreasonable
and way beyond the advised length of 20 – 25 pages. It is not only the examiner who has to wade
through all this verbiage; the candidate has to find the relevant evidence while working under pressure
in exam conditions. What is required is a brief documentation of evidence that they have done what
they were asked to do in the paper specification.
It is also worth considering the time spent by candidates on word processing so many pages of
unnecessary documentation; these hours could more usefully be spent on providing the evidence
actually asked for in the specification or on other topics required for this examination.
If a report goes onto a second page because of the number of entries, it is usually safe to include only
the first page of the report. If it goes onto a second page because of the number of columns, one
should look closely at the overall design and the fields included. The simple expedient of wrapping
the text in the column headings would often reduce the report width sufficiently to fit onto one page.
(See the points below on layout and contents of lists.)
There are still some candidates who number their pages in sections, e.g. design pages 1 – 5,
documentation pages 1 – 10, reports pages 1 – whatever. This is unhelpful to both the candidate and
the examiner. All pages should be numbered consecutively for ease of reference.
If the Practical Exercise Cover Sheet is not signed in future, it will not be assumed that the exercise is
the candidate’s own work and examiner’s will use their professional judgement for assessment
purposes.
Evidence
One examiner estimated that 30% of the scripts he looked at did not include the required reports, yet
nearly all of these had irrelevant ‘refinements’.
Report on the Examination Advanced Subsidiary & Advanced - Com
p
utin
g
￿￿￿
13
It is a fact that an examiner will mark some 300 plus scripts. They should be able to turn to each
reference and find the evidence clearly annotated on the pages identified by the candidate. The
instructions clearly say, for example, ‘Write Q 4 (c) (ii) in the margin, in the correct place, on that
page’. Too often, the candidate scribbled the question number anywhere on the page, bearing no
relation to the evidence he or she was indicating. Sometimes this did not matter as the evidence was
obvious on that page, but when searching for the row in a table, or some relevant code, the absence of
clear indication of where to look could result in that piece of evidence being overlooked and not
credited. As is the custom with this paper, there are some questions where the mark for the answer to
the question depends on satisfactory evidence of accomplishment.
It is in the candidate’s interest to be precise about referencing. Examiners look at the page
referenced. If the evidence goes onto the next page, candidates should indicate this in their page
reference.
Identification of output is important. One candidate offered a list headed ‘Daily list of remarks
completed’ as a list where a remark had caused a grade change. This was about to be rejected, but the
examiner happened to glance at the previous page where the query was displayed. It was clear that
this was, in fact, the daily list of remarks completed that had caused a change in grade, that is, it was a
perfectly valid reference in this case.
Hard copies
It has been emphasised in past examiner’s reports that a hard copy means just that, not a screen dump
inserted in a page of text. The brief says that ‘hard copies to prove the correct working of the system’
should be included. This year, it was sad to see so many marks missed because actual hard copies of
the required lists and the document to be returned to centres were not included in the documentation.
In some cases, there was even a comment in the documentation saying that ‘a print button is included
so that the list can be printed’.
It was acceptable that a hard copy be inserted onto a (full) page, so as to enable continuous page
numbering.
Layout & contents of lists
There being a number of lists produced in the solution to this exercise, it was disappointing to see
many poorly thought-out and poorly laid out lists. In particular, many lists included all fields,
whether relevant to that particular task or not. This even occurred with solutions produced in Access
where the list was the result of a query. For example, a list of records where a remark had changed
the grade did not need the date submitted, or whether the script had to be returned to centres. It is
hoped that candidates at this level would take care to produce relevant output from their systems.
In Excel solutions, too frequently a column width was determined by the length of the heading, rather
than the requirements of the data held in that column. Such candidates could have been recommended
to use text wrap. (Please note that spreadsheets will not be accepted as a method of solution for the
Practical Exercise from 2004 onwards.)
However, it is important to include relevant detail in documentation. It was expected that a list where
mark change affected grade should include old and new marks and old and new grades. The brief
actually listed what should be included in the document to be returned to centres, but not all
candidates included all the requested information. In particular, the ‘sentence to say there has been no
change’ was frequently omitted. In none of the chosen methods of solution was this difficult; its
omission therefore implies a lack of thoroughness in the execution of the task.
Computing - Advanced Subsidiary & Advanced Re
p
ort on the Examination
￿￿￿
14
Question 1
Apart from the obvious extra data items, such as original and remarked grades, for the remarking
process, acceptable items included the centre address for the return of the result and subject name.
Some candidates set up variables to record data such as whether the grade had changed or whether the
remark was overdue. These data, which were dependent on data already stored in the system actually
put the candidates at a disadvantage when it came to explaining exactly how their solution had
produced these lists, in question 2. Not accepted were data items irrelevant to the process of remarks
as described in the brief, such as a candidate’s data of birth or a candidate’s home address.
It was expected that the name given as the answer would agree with the name referred to in the
evidence. The ‘use’ implied that the field was necessary for the successful completion of the task;
‘Date_Rec is used to store the date the remark request was received’ was not creditworthy;
‘Date_Rec was used to produce the outstanding remarks report’ was.
An unqualified ‘number’ is not a sufficient data type for part b. Integer or text were expected for the
candidate number. Long Integer was not. Many candidates realised that the leading zero might be a
problem, and so chose text or an appropriately formatted integer. Appreciating that candidates whose
own candidate number had a leading zero would be more likely to think of that, the fact that a
candidate number was always a whole number was also credited.
Yes/No, binary, Boolean or a text field of length 1 gained a mark for part (b) (ii), and integer or byte
for iii.
Question 2
As explained above, marks were lost here because candidates did not point to the right type of
evidence or because the ‘hard copies’ were not, in fact, hard copies of the required list. Where the
question asked ‘Exactly how your solution produced a list’, it was not enough to say it looked at a
field called grade change, or out of date, without saying how that field contents was decided. A
significant number of candidates pointed to coding or a query which clearly used inappropriate
criteria, such as the centre number instead of a subject reference code. This implied that they could
not ‘read’ the coding themselves. It does emphasise the importance of documentation being clear and
coding annotated.
In (a) (ii), a hard copy list of remarks where grades had changed as asked for. A list which clearly
showed some grades unchanged, or where some re-mark grades had not been entered, suggesting the
remark had not been completed, did not provide sufficient evidence.
In 2 (b), an algorithm was asked for. Candidates should already have worked out how their solution
could produce a daily list of remarks where the candidate’s grade had been affected by a mark change.
They were now asked to write an algorithm to produce a list of candidates whose grade had not
changed. Examiners were fairly lenient in marking these, accepting pseudocode, coding, or a list of
steps. SQL was accepted, although that does not actually produce an algorithm. A QBE grid was not
credited, although an expression within it could gain a mark.
The required steps were:
1. To select all records. For each:
2. To check the remark had been completed that day
3. To compare the old grade with the remark grade
4. If the grade was unchanged, then to print that record
5. To loop back until all records had been examined.
Report on the Examination Advanced Subsidiary & Advanced - Com
p
utin
g
￿￿￿
15
Question 3
‘I loaded up the AQA website and found the page with the UKAB logo on it. Then using CTRL & prt
scr buttons I took a screen shot. I then loaded up paint and copied the screen shot into paint. Then I
cropped the screen shot until I had the UKAB logo on its own. I then copied the logo and pasted it
into Word.’ This is an example of a good answer to question 3 (a). Many candidates did not gain
these marks either because their ‘hard copy’ was not a hard copy, or, all too frequently, because their
‘document to be returned to centres’ contained candidates from more than one centre. Where
candidates had included more than one of these documents, it was hoped that at least one would
include the sentence saying that there had been no grade change, as specified in the brief. For a
significant number of candidates, this was not the case.
It was accepted that, in some programming languages, it was difficult to include a scanned image of
the UKAB logo. Many centres had raised this point with the Principal Examiner prior to the
examination and it had been suggested that pre-printing the logo, so producing ‘headed notepaper,
would be an acceptable solution. Other candidates produced their own version of the logo. This was
credited if an explanation was given as to why it was not possible to use the one in the brief.
Although this is a computing paper, it is regrettable that many of the documents to be returned to
centres demonstrated a poor level of command of English. One candidate ‘regretted any convenience
(sic) that may have been caused’ by the need for a remark. Capital letters frequently occurred in the
wrong place.
In part 3 (b), many candidates did not appear to understand the term ‘criteria’. A two-mark answer
might be ‘The text should be easy to read, so I used Times New Roman 12 point.’ Candidates had
arranged the information in a table or chosen landscape so that all the information should fit onto one
line. Important text was highlighted and important information included.
Candidates who use application packages should realise that buttons and drop down lists do not
appear on a document!
A significant number of candidates sensibly suggested that they had included the centre address so
that they could use window envelopes to send the document to centres. It was therefore unfortunate
that most of them then placed their ‘centre address’ top right instead of slightly lower on the left.
Question 4
In part 4 (a), candidates who vaguely referred to a validation rule to prevent a mark greater than 100
were not credited. ‘” I placed a validation rule on both mark fields that said ‘ >=0 And <=100’” was a
good clear answer.
Candidate numbers are unique within centres but not between centres. It was particularly
disappointing in this question how many candidates with an Access solution ‘got round’ the problem
of duplicate candidate numbers by using autonumber to create another primary key. The use of
autonumber is rarely the best way of creating a primary key in an Access table.
For this problem, a composite key was the expected solution for an Access approach, with a
concatenated key being appropriate for any method. This clearly ‘allowed’ two candidates in
different centres to have the same candidate number; as specified in the Brief. A separate sheet, array
or table for each centre was not accepted. As the brief had not said that a candidate might ask for a
remark in more than one subject, this facility was not expected. However, candidates who had
allowed for this were in no way penalised. The brief clearly stated that testing should cover
situations in which marks are both increased and decreased. Some candidates referenced a test plan
for 4 (c), but rarely did it say ‘I will increase the mark to check that the grade changes’.
Computing - Advanced Subsidiary & Advanced Re
p
ort on the Examination
￿￿￿
16
Usually it tested extreme or negative values. Many candidates still appear not to know the concept of
‘test data’. Answers such as ‘mark (or ‘grade’) increased’,’ mark decreased’, are not test data,
although evidence was credited if clearly referenced and clearly showing both the original and the re-
marked mark.
Question 5
The format of the Entity-Relationship diagrams was usually correct, and most candidates understood
that one centre entered many candidates. In real life, each candidate may take many subjects. Thus
many candidates take many subjects. It was also accepted for this system that many candidates take
one subject. Finally, many centres enter candidates for many subjects.
In part (b), many candidates ignored the fact that the analyst had completed his research in the past
before their solution was created. He could not have examined ‘the current system’, or ‘the new
system’, or ‘the specification’ to find what was required.
Answers had to be in context to be credited. It was expected that a systems analyst might have
interviewed an official at the Board to find which reports were required from the system, and
observed the old system, to find out the expected volume of data. Questionnaires are used when
collating information from a number of sources, so while a questionnaire to Centres might be
appropriate, a questionnaire to the Board was not.
Phrases like ‘Look at’, ‘Look for’, ‘talk to’ will not gain marks when describing the methods an
analyst uses
Question 6
Part 6(a) was looking for fairly standard methods of data validation or verification and most
candidates gained at least one mark. However, the question offered two marks for a full answer.
Those candidates who were sufficiently familiar with the term verification to use it correctly as in
‘Type the data in once and then type it in again in verification mode’, or ‘ a second person types it in
verification mode’ would gain the second mark.
In part 6(b), not all candidates appreciated that the security techniques were to be particularly aimed
at unauthorised access. Thus regular back-ups were not actually appropriate. Firewalls, security and
encryption were the most common creditable responses. Marks were lost by those candidates who,
despite the fact that they were already given in the stem, suggested password protecting the files and
using access rights.
Question 7
Evaluation and maintenance is another area which many candidates do not appear to understand fully.
Factors which would be looked at in an evaluation include a comparison with the original objectives,
the number of complaints and of breakdowns or maintenance incidents and ease of use. General
statements about efficiency were not sufficient for a mark but comparisons of the cost, or the
turnaround time of the new system with the old, were.
An effective system should be able to alter grade boundaries, add extra subjects and archive or delete
old data, without needing ‘maintenance’. The changing of the format of the centre or candidate
numbers was one reasonable suggestion which came up a few times. Candidates suggested the
correction of a fault that had been discovered after the system had been running for a while.
Alternatively, new technology or new developments in software might be a cause.
The list of items of documentation which would be necessary for effective maintenance is a long one,
but does not include a list of candidates or of centres. Coding should be annotated.
Report on the Examination Advanced Subsidiary & Advanced - Com
p
utin
g
￿￿￿
17
CPT 4: Processing and Programming Techniques
Two topics were new to this year’s specification for Module 4: the representation of negative numbers
by Twos Complement, in the section on Data Representation in Computers, and Process / Task /
Management in the Operating System section. Both topics were examined in a small way this time.
The Twos Complement question was answered correctly by about 50% of the candidates, but, sadly,
candidates were less successful with the question on threads.
Much of the rest of the paper re-visited topics which have appeared in past papers, although
sometimes with a slightly different approach from recent exams. Thus candidates who had covered
the whole specification, revised well and read the questions carefully, so that their answers were
appropriate, scored well, especially if they were familiar with the relevant terminology. Techniques
such as the vectored interrupt mechanism, (question 2) and paging (question 7 (b)) are simply
bookwork, as are the correct definition of virtual memory (question 7 (a)), an instruction set and the
different addressing modes (question 9).
Too many marks were missed by candidates’ careless use of terminology. A memory location is not
the same as an address, (question 9(c)), if main memory is meant it needs to be stated thus, (question
7), and the number 11111111 is interpreted as 11 million, one hundred and eleven thousand, one
hundred and eleven, unless it is followed by the subscript 2, (question 9 (e)) .
Question 1
Many candidates failed to score as highly as they should have through their lack of ability ‘to organise
relevant information clearly and coherently’. The basic argument is that the fact that a queue is a
‘First In, First Out ’ structure, while a stack is a ‘Last In, First Out’ structure. This means that if the
elements of a queue are pushed onto a stack, their order, when pushed off the stack, is reversed.
Incidentally, the word is ‘queue’, not ‘que’ or ‘cue’ – mistakes prevalent in candidates’ answers.
Question 2
With many candidates this was a classic example of ‘Answer the question posed, do not give the
answer to the question on this topic from the past paper which you did last week’. This question did
not ask for a definition of an interrupt, nor how interrupts are prioritised. It asked for a description of
how the vectored interrupt mechanism works; that the interrupting device supplies an offset which is
added to a base address to give the start address of the interrupt service routine.
The advantages of this mechanism are that new interrupt service routines can be easily introduced,
and ISRs in general can be dynamically loaded or are re-locatable.
Question 3
Some candidates simply re-wrote the question. Others gave quite commendable inheritance diagrams,
which, as they were not asked for, gained no marks. Examiners were lenient in their assessment here,
giving marks for genuine attempts at coding.
Question 4
Most candidates recognised that the Twos Complement number in part 4 (a) was positive, although
the estimate of its size, which depended on their realising that the exponent was negative, was often
wildly out. Where candidates failed to convert the second number correctly, credit was given for
relevant working. The commonest error was in not appreciating that the binary point originates
Computing - Advanced Subsidiary & Advanced Re
p
ort on the Examination
￿￿￿
18
between the two leftmost digits. A less common, but more dispiriting error, was the use of the
exponent as a power of 10.
Processing numbers in fixed point is quicker than in floating point, less processing is required, but
‘calculation is quicker’ was insufficient, and ‘easier to work out’ seemed to show a complete lack of
understanding of human computer interaction. A number of candidates said it was easier to
understand but this was not an accepted answer. Fixed point representation can give greater accuracy
or precision, although many candidates thought otherwise. Thus fixed point representation would be
used where maximum precision is required or where the possible range of numbers is small, or of a
set format, such as with currency.
Question 5
5 (a) was reasonably answered by most candidates, with waiting for input or output, being interrupted
by a higher priority process such as an interrupt, or the end of a time slice being frequently offered.
For part (b), the candidate who wrote ‘This is not in the syllabus’ must have been using an out of date
specification. Very few candidates had any understanding that threads share more of their
environment with each other than do processes under multi-tasking. They often share a single address
space and set of global variables, and so are distinguished only by their program counters and stack
pointers. This minimises the time required to switch threads.
Question 6
6 (a) (i) and (a) (iii) were answered correctly by most candidates as Mon and False respectively, (not
[Mon]). However, in (a) (ii) [Head(Days)} is the list [Sun] and the tail of this list is the empty set [ ].
Had these data been stored in a one dimensional array, instead of in a list, each element could have
been accessed directly using the subscript, rather than having to be found through a sequential search
or using the Head / Tail functions defined.
Question 7
Many candidates had a good understanding of the technique of virtual memory, but a common flawed
answer was that it ‘made the computer think it had more memory than it actually had’. Marks were
lost here by careless use of the term memory to mean main memory. A good answer said that the
hard disk is used to supplement main memory when it is not large enough for the execution of a
process. Many candidates used the word ‘data’ when they should have written ‘program’ or
‘program and data’. This implies that they do not really understand that the program is also swapped
in and out of main memory.
Candidates were less able to explain paging precisely enough to gain all the available marks. Main
memory and the virtual address space of the process are divided into pages, of the same fixed sized,
which are loaded into main memory as required and then copied back to the disk before being
overwritten. The Page Management Table stores which pages are loaded and where.
Question 8
This was well answered by many candidates, although here again, marks were lost through lack of
care in carrying out the clearly written instructions. The instructions in part (a) were to circle and
label the three parts of the tree. An un-circled, or (more rarely) unlabelled part was not credited.
Very few candidates were unable to draw the left and right sub-trees correctly, and most could
perform in-order traversal.
Report on the Examination Advanced Subsidiary & Advanced - Com
p
utin
g
￿￿￿
19
Question 9
Although many candidates scored well on this question, others showed a basic lack of understanding
of this topic area. For example, part (b) asked ‘With 6 bits of the op code reserved to denote basic
machine operations, how many basic machine operations may be coded?’ Incorrect answers included
1, 2, 6, 13, and 63.
A significant number of candidates could not describe the addressing modes correctly, although
whether it was their lack of understanding of this topic area or more general poor written
communication skills was not always clear. A memory location is not an address. Operand fields do
not hold instructions or the memory address of instructions.
Many candidates scored well in part (d), although some did not take advantage of the fact that the first
instruction had been given, so providing a good clue as to what was expected and in what format.
A significant number of candidates completely mis-interpreted part (e), giving B2 and B3, or their
decimal equivalent, as their answers. There was confusion as to the number of possibilities offered by
a given number of bits, 2
n
as required in (b), and the highest address offered by a given number of
bits, 2
n
-1, as required in (e).
Question 10
Most candidates scored well in question 10, with some candidates earning 50% of their total mark in
this question. The interpretation of the given facts and rules was well answered. In part (c) the
examiners were looking for the continuation of that set of facts and rules, so a good solution added a
new noun phrase rule for an adjective and a noun, and then incorporated this into a new sentence
phrase rule. IF and AND were expected to be in upper case, as were all parameters. Descriptors
should be completely in lower case.
Candidates should appreciate that legible handwriting, reasonable spelling, punctuation and grammar,
and the appropriate use of specialist vocabulary can all have an impact on their final mark. An
examiner cannot mark as correct a term he or she cannot read. In many questions at this level, credit
will only be given for the correct use of the appropriate terminology. The ‘Information’ section on
the front of every script should be studied carefully and examination technique amended accordingly
to meet requirements.
Computing - Advanced Subsidiary & Advanced Re
p
ort on the Examination
￿￿￿
20
CPT 5: Advanced Systems Development
General
Pleasingly, many candidates showed good knowledge of the networking content of the specification
examined in question 9. Also pleasingly, candidates responded well to question 6 which was based
partly on contemporary devices and comprehension of their uses. Candidates responded less well to
question 7 which required some careful thought and use of imagination on the part of the candidate.
The answers given by many candidates were often not benefits, or, if they were, not sufficiently well
expressed to gain credit or were wrongly targeted, i.e. an answer for a benefit to a pharmacist was
actually a benefit to a patient. The Collins English dictionary states that a benefit is something that
“improves”. Candidates should practise answering this type of question and be provided with
opportunities to develop their critical thinking. Only the better candidates were able to offer credit
worthy reasons for part 7(d), which required an understanding of why large-scale software systems
fail. A similar situation arose in Q4(c) where answers of many candidates revealed a limited
understanding of computer systems. Whilst it is unrealistic to expect candidates to have direct
experience of scenarios similar to the one examined in question 4, it is possible for candidates to gain
understanding and insight indirectly through exposure to case studies.
Disappointingly, many candidates’ knowledge of SQL was imperfect. Many candidates also seemed
unable to accept that a many-to-many relationship could exist between two entities. Many of these
candidates included additional entities so that one-to-many relationship links could be drawn.
Significant numbers of candidates were unable to identify correctly the diagram in question 3. Typical
responses were data flow diagram or entity-relationship diagram. Many candidates, given that a
MICR reader was used in the diagram to read payment stubs, were unable to expand the acronym
MICR. Many candidates answered that the encoding method used was encryption.
Question 1
The majority of the candidature was able to offer two credit-worthy fact-finding techniques from
amongst the list: interview, observation, survey and examination of paperwork.
Question 2
Few candidates were familiar with the key-to-disk method of data input. Many simply answered that
the data would be entered through the keyboard, full stop. Implicit in such an answer was that
processing took place in real-time. The reality is that it does not as there would be little control over
the propagation of transcription errors. If this did happen, as soon as data is processed it could be used
elsewhere in the system even if it is incorrect. In a typical data processing scenario, involving entry of
data from a large number of paper forms, the data would be keyed in by a clerk and stored on disk
prior to processing at a later time. This enables checking of the data to be carried out.
Many more candidates gained credit for answering that the data on the forms could be scanned in and
processed using the techniques OCR or OMR.
In the case of both key-to-disk and OCR/OMR many candidates were able to offer acceptable
justifications. A very common answer was that often handwritten forms required some interpretation
and therefore key-to-disk was appropriate. In the case of OCR/OMR, a very popular answer was data
could be entered very quickly if the process was automated.
Report on the Examination Advanced Subsidiary & Advanced - Com
p
utin
g
￿￿￿
21
In part (b), verification of the data reduced data entry errors was a very popular correct answer. A
well-designed form with boxes for where handwritten responses should be written and use of capital
letters was a popular correct response and could be used for OCR processing.
Question 3
Those candidates familiar with system flowcharts scored well on this question. The use of
emboldened words in the stem aided candidates in choosing the correct labelling for the components
of the systems flowchart. Disappointingly, many candidates either omitted to add to the diagram the
data flow between the process box and the on-line storage symbol, or chose, incorrectly, to make it
unidirectional. The difficulty that several candidates experienced naming this diagram and the
encoding method has already been referred to in the introduction. The encoding method is Magnetic
Ink Character Recognition, hence the use of a MICR reader.
Question 4
Several candidates failed to gain credit for their answer for parallel processing because they omitted
to give some indication that the operation of the old system alongside the new is time-limited, i.e.
once new system proved the old system is abandoned.
Too often candidates’ explanation for phased conversion fitted an explanation of pilot conversion.
Essentially, the examiners were looking for parts of old system gradually replaced by new system in
stages. Several candidates answered, without gaining credit, that phased conversion meant replacing
the old system department by department in stages.
Candidates responded successfully to part (b), on the whole, with training and copying data from the
old system to the new being popular answers.
Candidates were less successful answering part (c) with many candidates posing general questions
rather than specific questions. For example, one response simply gave “How maintainable is the
system?” as the answer. The developers should be asked questions relating to how long will it take to
correct errors in the software, whether upgrades will be available and how long will it take to upgrade,
whether it is possible to change parameters such as the rate of VAT, the number of users, et cetera.
Such responses show real understanding or appreciation of the maintenance phase of a software
system. For example, a subscription to virus checking software usually includes regular updates which
are available on-line and are semi-automatic.
Question 5
Part (a) was well answered on the whole with barcode and magnetic stripe being the most popular
answers. A smart card was another acceptable answer. Some candidates substituted microchip for
smart card which was accepted, but chip was not.
Few candidates obtained both marks for part (b) because either they failed to state that biometric
information was stored on the card or that a comparison was made with biometric information stored
in a central database.
This question and several other questions demonstrated a tendency amongst the weaker candidates to
supply answers that left much unsaid. Part (a) asked for a description but several candidates simply
wrote one-word answers or acronyms such as MICR. Apart from demonstrating poor examination
technique, this could be construed as evidence that the candidate’s thinking and communication skills
were underdeveloped.
Computing - Advanced Subsidiary & Advanced Re
p
ort on the Examination
￿￿￿
22
Question 6
Candidates performed well on most of this question. Part (b)(i) was less well answered across the
candidature. If the CODEC has its own hardware then the loading on the main processor can be
reduced. Consequently, compression and decompression of videos will be faster.
Question 7
Many candidates failed to gain credit on several parts of this question because candidates were often
unable to think carefully enough about the given scenario and unable to differentiate a statement of
fact from a statement of a benefit. For example, several candidates responded to part (b)(i) with the
answer which gained no credit, “ a patient can book an appointment to see a doctor by using the
system”. The more astute candidates gained credit for answering “a patient can save time by being
able to view appointment slots on-line to book a slot directly thus avoiding waiting for a receptionist
to answer the phone”. It is a fact that a patient, while on holiday, could go to any doctor in the country
if taken ill and be treated. However, if every doctor’s surgery is connected to the networked NHS then
the benefit to the patient is that the patient is more likely to get treated safely if the doctor has access
to the patient’s medical history or treated more quickly because the doctor will not have to ask quite
so many questions. Many candidates stated the fact that applies without a networked NHS. However,
fewer candidates managed to state the benefit. Candidates need opportunities to develop thinking and
communication skills that are necessary for success in answering this type of question.
Candidates performed much better on parts (a) and (c) which tested bookwork learning. However, it
was evident that candidates from whole centres were not familiar with the three level architecture of a
DBMS or how a DBMS handles concurrent access to data.
Only the better candidates performed well on part (d). Many candidates demonstrated a poor
understanding of the software/system development process and reasons for its failure or why projects
are not finished on time. Candidates tackling Computing coursework should have gained practical
experience of the analysis phase of systems development and the formal methods used to document
this phase. In particular, the gathering and documenting of user requirements. Often systems do not
perform as users expect because the users’ requirements are poorly specified. A system is built to
specifications and if these are imprecise, incomplete, inadequate or incorrect the users will not get
what they expect. Other reasons are that the hardware is inadequate or the setting up/configuring of
the system has not been done properly or the system is not being run properly because of a lack of
trained personnel.
Systems do not work at all because they contain software errors or the design is faulty or the system
was not tested adequately and for several other reasons.
Part (d)(iii) required a candidate to answer with a little more subtlety than in the previous parts to (d).
Failure to complete on time is not caused by debugging errors in the software or the time taken to
transfer data into the new system but in underestimating the time taken to do these things.
Potential problems are not foreseen because the existing system is not properly analysed. Projects that
are poorly managed tend to overrun their deadline. It is surprising that many candidates were not
unable to map their own experiences of their Computing coursework onto this scenario.
Question 8
Candidates this year did not perform as well as in previous years on this type of question. The greater
emphasis on writing SQL exposed gaps in some candidates’ knowledge. Whilst many candidates
coped with a single table query, fewer were able to give the correct syntax for a query involving the
joining of two tables. This question also exposed a lack of formal treatment of the theory of SQL.
Several candidates referenced the tables Book and BookCopy as tblBook and tblBookCopy,
Report on the Examination Advanced Subsidiary & Advanced - Com
p
utin
g
￿￿￿
23
possibly because they have experienced a software package that creates tables with this notation.
Candidates lost marks for using quotes on the accession number 1234 and not using quotes on the
ISBN value of “1-57820-082-2”.
In part (a) (ii) several candidates could not cope with the many-to-many relationship between entity
Book and entity BookCopy and proceeded to add intersection entities so that two or more one-to-
many relationships could be drawn, not accepting that the many-to-many relationship entity still
existed. Other candidates drew an incorrect relationship. Many candidates successfully identified that
the relationship between Borrower and BookLoan was one-to-many.
Many candidates correctly gave mail merge as the process in part (c).
Question 9
Most candidates made a reasonably successful attempt at this question but router was not well
understood. Very few candidates knew that routers use the IP addresses of packets to route a packet
across the Internet. They forward data packets if the destination is not on the current network. Routers
are at the core of the Internet. Without routers we would not have an Internet. The specification
states: “define this term and consider when and why routers are used. In particular, consider how
routing is achieved across the Internet”. Understanding what is a router and routing is as fundamental
to internet working as an understanding of what a nucleus is, and that splitting the nucleus produces a
vast amount of energy, is to nuclear physics.
The better candidates made a reasonable attempt at part (c). A socket address is considered to be IP
address of the host plus port number. A port number connects an application running on a host to the
network. Web servers listen on port 80 (or 8080) so a web client, a browser, must send a request for a
web page to this port on the server. Of course, the web server host must be located first. This is
achieved by using the web server host’s IP address. An analogy is dialling the number for an
organisation (IP address) and then when connected dialling the extension of the person within the
organisation (port no) to be contacted.
A client browser must first connect to a web server before sending a get web page request. The web
server creates a new socket to deal with each connection which to all intents and purposes is
equivalent to allocating an unused port to handle the new connection although web servers are written
to handle all connections through port 80. Several candidates were able to describe this mechanism
sufficiently well and therefore gained credit.
Computing - Advanced Subsidiary & Advanced Re
p
ort on the Examination
￿￿￿
24
CPT 6: Coursework
General
This report provides general feedback on the overall quality of project work for A level computing
candidates. In addition, all centres receive specific feedback from their moderator in the form of a
short report that is returned after moderation. This reporting provides an ongoing dialogue with
centres giving valuable pointers to the perceived strengths and weaknesses of the projects moderated.
Nearly all projects are returned to centres after moderation. A few are first copied and then retained
by the Board as archive materials but centres are always notified in this circumstance.
Administration
Centres are reminded that each project must have the following attached to the front.
• A Candidate Record Form showing the marks awarded and the declaration that the project
is the candidate's own work;
• A Project Log Sheet. This provides an outline of the project, completed by the candidate
and agreed by the teacher, and a commentary in sections C, Project Development, and D,
Teacher’s Notes, as to the reasons for marks awarded. Some candidates did not state the
software used on their project log sheet. This is essential as sometimes it is not clear from
the evidence provided in the report.
The candidate and the assessor must sign both the Project Log Sheet and the Candidate
Record Form. The latter indicates the authenticity of the project.
The moderation team thanks all those teachers who took the trouble to make comments on the Project
Log Sheets or provided a centre-devised marking commentary. Perceptive comments from a teacher
really do aid the process of moderation. Again, moderators were concerned about the number of
incorrect additions entered on the Candidate Record Forms. Please can centres check their arithmetic
as we can only check the projects sampled.
Centres are reminded that their Centre Mark Sheets must arrive with their moderator on or before the
specified date. Centres are reminded that projects should be forwarded in the board-provided sack.
Candidates should wherever possible provide a single project report, not a set of multiple folders as
these can easily be parted from each other and do not aid the process of moderation. Candidates
should not include floppy disks or CD-R/RWs with their project reports, as the moderators do not
sample these items. Work should not be posted by registered post or any another carrier that requires
the parcel to be signed for.
Software
The majority of projects were completed using MS ACCESS combined with Visual BASIC, and, for
an increasing number, combined\with Delphi. Programming languages used included Turbo Pascal,
Delphi, Visual BASIC, C/C
++
, QBASIC, RM BASIC, ASP, PHP and Java.
Project Development
The majority of candidates developed a database system but this is only to be expected when
ACCESS is the most popular software development tool. Candidates should be reminded of the
specification requirement for processing of data and the use of a programming language. 'Data Entry
Only' systems, Web pages just providing information and code free implementations were still
submitted by some candidates as suitable projects. These types of projects limit the marks available
to the candidate.
Report on the Examination Advanced Subsidiary & Advanced - Com
p
utin
g
￿￿￿
25
The range of projects was not as diverse as seen previous years. However, there were still the same
number of libraries, newsagents and video shops. Whilst appreciating it is sometimes difficult to
ensure that all students choose demanding topics, or to motivate them, it should be the role of the
teacher to ensure that the work completed meets the project specifications and that appropriate
software is selected.
Project Reports
The work of the candidate is judged on the project report submitted; therefore candidates must include
evidence for all sections in order to justify the marks awarded. These reports should be word-
processed and set out in the recommended sections with a contents page at the front. It is assumed
that ‘A’ level candidates have a good knowledge of the facilities offered by a word-processing
package. It would be appropriate to demonstrate these in their report including, at a minimum, use of
headers, footers, document sections, automatic page numbering, spellchecker, grammar checker, a
style sheet and other tools as appropriate.
Again, centres are reminded that in a number of sections only samples of evidence are required for the
report. Many candidates are including a full set of evidence instead of a sample considered in depth,
e.g. a full set of I/O designs quickly sketched rather than a carefully selected representative sample
clearly annotated as to purpose, underlying data and HCI rationale. Extra marks are not available for
the volume of evidence. It is the quality that should be assessed.
Analysis
This section provides the starting point for the candidate's project and requires full and careful
consideration by the candidate. A thorough investigation using recognised methodology will lead to
realistic, specific objectives that can be addressed by the candidate. However, in order to determine
these objectives the evidence from this investigation must be analysed by the candidate not just
reported.
Candidates also need to set the system bounds for their proposed system as an in depth analysis often
reveals a larger system than the candidate has time to develop to the appropriate depth and required
complexity to meet fully both their user’s needs and the requirements of the GCE ‘A’ Level
specification.
Common problems include:
• Interviews or questionnaires used as the sole means of investigation. Other methods can reveal
valuable information, e.g. document search, observation etc.
• Objectives stated without supporting evidence.
• Objectives that are too general: these need to be specific and measurable and clearly related to the
candidates proposed system. A few vague statements are not sufficient.
• Candidates not using a recognised methodology for their Data Flow diagrams.
• Feasibility of potential solutions not included in this section but left to the design section.
Design
This section provides evidence for the design of the solution not the implementation. Many
candidates simply provided evidence from their implementation rather than showing the rationale for
their design. It may be appropriate to include some I/O screens from the implementation but in order
to gain credit they must be clearly annotated as to the HCI design.
Computing - Advanced Subsidiary & Advanced Re
p
ort on the Examination
￿￿￿
26
The expected contents for this section are set out in the specification but the elements must be applied
to the candidate's project rather than considered in a theoretical manner. Also the specification clearly
identifies that evidence needs to be in the form of representative samples. Candidates will not gain
extra credit for including a full set of evidence, e.g. only a sample of the validation checks and their
rationale is required.
Common problems include:
• Elements of design missing, e.g. system overview, security, design of processing algorithms.
• E-R model, but no normalisation or incorrect normalisation. Evidence for normalisation provided
as a relationship diagram printed out from ACCESS; this is evidence of implementation not
design.
• Superficial consideration of validation or evidence provided from implementation on how the
validation was performed, not why it was required.
• Theory of testing rather than a strategy applied specifically to the project.
• Inclusion of a work schedule is not required.
Technical Solution
Candidates should provide clear evidence of writing their own code in order to gain good marks in
this section. Much of the evidence for this section is gained from the Systems Maintenance section of
the candidate’s report.
Better candidates showed the use of well structured programming techniques that included the use of
local variables in procedures and functions together with parameter passing, the use of structured
naming conventions. Very few candidates made good use of object-oriented programming.
In order to provide evidence for this section, candidates should structure their listings carefully, use
meaningful variable/procedure/function names and use the commenting facilities provided by the
language. Candidates using package-driven features are reminded that evidence of tailoring needs to
be provided. This includes printouts of macros, queries, and relationship diagrams. Also lists of
forms, reports, tables, buttons, macros with their name and use; evidence of form and report design
together with implementation of validation checks and evidence of processing of data, e.g. the
automatic updating of a student table once an examination has been passed.
Common problems include:
• No evidence of the tailoring that has taken place.
• Forms, reports, queries etc not presented in ‘design view’ or annotated by the candidate.
• Evidence of corrective action; this is not required neither is an implementation commentary.
• SQL listing when QBE has been used; an annotated printout of the QBE grid would be more
useful.
• Listing of evidence that has been produced by wizards; only a mention of the items is required, as
the candidate cannot claim evidence for the work done by the wizard!
• Programs that have been automatically produced claimed as written by the candidate.
System Testing
Most candidates provided a test plan but this should test the actual system as detailed in the design
and implementation sections and conform to the candidate's test strategy. Candidates are reminded of
the requirement to show evidence of processing with files/tables being updated by their system. The
Report on the Examination Advanced Subsidiary & Advanced - Com
p
utin
g
￿￿￿
27
use of erroneous and extreme data in order to test validation checks must be included. It is not
sufficient to provide a summary of the testing; there must be a test plan that is cross-referenced to
sample evidence of actual testing with screen dumps clearly annotated to show the result of the test.
Common problems include:
• No hard copy evidence of testing provided.
• Two copies of the test plan one with actual results and one with expected results; only one is
required and it must show how the actual results match the expected results, or if not what the
problem was.
• Evidence for all tests; only a representative sample is required.
• Much testing of the use of buttons and little testing of validation rules and processes.
System Maintenance
The marks for this section are for the way the system is documented not what the system does, so a
candidate can gain marks in a higher performance level for their System Maintenance than their
Technical Solution.
Evidence provided for this section varied in both quantity and quality from centre to centre, showing
that candidates do need guidance from their teachers as to the expected contents. Many candidates
scored their lowest performance level marks in this section either running out of time or not having
appropriate evidence to include. Conversely some centres appeared to award high marks for very
little appropriate evidence.
Here is the list of elements that candidates need to include in order to obtain good marks:
• an overview of the system developed;
• samples of detailed algorithm design (i.e. one or two processes) using recognised methodology,
e.g. pseudo-code;
• full annotated listings of program code including macros and SQL, these listings could be in an
appendix but need to be clearly referenced from the report;
• procedure and variable lists/descriptions for programs or details of forms, reports, tables, buttons,
queries for packages.
Common problems include:
• No system overview.
• No algorithms, possibly because little processing has taken place.
• Reference to the design section stating that elements were as described there, rather than
providing evidence from the implementation.
• Details of how to tailor the package included rather than details of how the package was tailored,
e.g. ‘How to set up a query’ rather than evidence of queries produced by the candidate. Yet again,
some candidates produced implementation commentaries with their report to show how they had
developed the solution. This is in the style of A level ICT projects and very often cannot be used
as System Maintenance because it is ‘how to use the package’ not to maintain their system.
• Queries, forms, reports etc not shown in design view and not annotated by the candidate.
Computing - Advanced Subsidiary & Advanced Re
p
ort on the Examination
￿￿￿
28
User Manual
Most candidates still did not follow the instructions given in the specification to produce a sample of
the User Manual; instead they produced a full report, very often professional in its quality, but
missing out on elements that should be present for the award of high marks. The elements required
are: an introduction, installation procedures, how to a use a section of the system together with
troubleshooting and error recovery procedures for that section.
Common problems include:
• Screen shots that cannot be read.
• Instructions without sample screen displays, or screen displays without any data.
• No details of any output other than screen displays.
• No troubleshooting and error sections.
• Network paths given to start up programs; candidates not creating executable files for programs or
informing their user they need a copy of the application package before using their solution.
Appraisal
This section is frequently a superficial afterthought and is sometimes missing from candidates'
reports. Few candidates use their remarks in this section to provide a critical appraisal of the system
objectives set out as the result of their analysis. Candidates are reminded that improvements and
enhancements need to be specific and supported by evidence of how the candidate would propose to
achieve them. Constructive and effective end user feedback may be difficult to obtain but in order to
add to the statements provided by the candidate, it needs to be more specific than a thank you letter on
headed notepaper. Centres are again reminded of the need to confirm that this feedback is genuine.
Common problems include:
• Objectives restated as being met, rather than giving details of how those objectives were met/not
met/partially by the system developed.
• Candidates using this section to chastise themselves for their own shortcomings, rather than
focusing on their system.
• Obviously invented user feedback.
Quality of Communication
Most candidates provided a well-structured report divided into appropriate sections. However, there
were many candidates who did not number their pages so the cross-referencing needed for testing was
not possible. Also many candidates provided evidence in appendices to which no reference was made
in the main body of the report.
Report on the Examination Advanced Subsidiary & Advanced - Com
p
utin
g
￿￿￿
29
Mark Ranges and Award of Grades
Unit/Component
Maximum
Mark
(Raw)
Maximum
Mark
(Scaled)
Mean
Mark
(Scaled)
Standard
Deviation
(Scaled)
CPT1
65
65
35.4
10.8
CPT2
65
65
29.0
10.2
CPT3
65
65
22.6
11.0
CPT4
65
65
25.7
9.5
CPT5
65
65
29.1
8.5
CPT6
60
60
37.5
11.6
For units which contain only one component, scaled marks are the same as raw marks.
CPT1 (10589 candidates)
Grade
Max.
mar
k
A
B
C
D
E
Scaled Boundary Mark
65
48
43
38
33
29
Uniform Boundary Mark
105
84
74
63
53
42
CPT2 (11984 candidates)
Grade
Max.
mar
k
A
B
C
D
E
Scaled Boundary Mark
65
40
35
30
26
22
Uniform Boundary Mark
105
84
74
63
53
42
CPT3 (11795 candidates)
Grade
Max.
mar
k
A
B
C
D
E
Scaled Boundary Mark
65
35
30
25
21
17
Uniform Boundary Mark
90
72
63
54
45
36
CPT4 (5963 candidates)
Grade
Max.
mar
k
A
B
C
D
E
Scaled Boundary Mark
65
38
34
30
27
24
Uniform Boundary Mark
90
72
63
54
45
36
Computing - Advanced Subsidiary & Advanced Re
p
ort on the Examination
￿￿￿
30
CPT5 (6347 candidates)
Grade
Max.
mar
k
A
B
C
D
E
Scaled Boundary Mark
65
40
36
33
30
27
Uniform Boundary Mark
90
72
63
54
45
36
CPT6 (6823 candidates)
Grade
Max.
mar
k
A
B
C
D
E
Scaled Boundary Mark
60
49
44
39
34
29
Uniform Boundary Mark
120
96
84
72
60
48
Advanced Subsidiary award
Provisional statistics for the award (10203 candidates)
A B C D E
Cumulative %
12.3 25.7 42.7 60.4 76.6
Advanced award
Provisional statistics for the award (7041 candidates)
A B C D E
Cumulative %
12.7 29.8 50.9 72.6 90.0
Definitions
Boundary Mark: the minimum mark required by a candidate to qualify for a given grade.
Mean Mark: is the sum of all candidates’ marks divided by the number of candidates. In order to
compare mean marks for different components, the mean mark (scaled) should be expressed as a
percentage of the maximum mark (scaled).
Standard Deviation: a measure of the spread of candidates’ marks. In most components,
approximately two-thirds of all candidates lie in a range of plus or minus one standard deviation from
the mean, and approximately 95% of all candidates lie in a range of plus or minus two standard
deviations from the mean. In order to compare the standard deviations for different components, the
standard deviation (scaled) should be expressed as a percentage of the maximum mark (scaled).
Uniform Mark: a score on a standard scale which indicates a candidate’s performance. The lowest
uniform mark for grade A is always 80% of the maximum uniform mark for the unit, similarly grade
B is 70%, grade C is 60%, grade D is 50% and grade E is 40%. A candidate’s total scaled mark for
each unit is converted to a uniform mark and the uniform marks for the units which count towards the
AS or A-level qualification are added in order to determine the candidate’s overall grade.