GCE Computing. COMP3 Problem Solving, Programming, Operating Systems, Databases and Networking Summer Version: 1.0

Transcription

1 GCE Computing COMP3 Problem Solving, Programming, Operating Systems, Databases and Networking 2510 Summer 2016 Version: 1.0

2 It was pleasing to see the quality of responses to some questions and in particular the improvement to responses to question parts on topics that have been asked about previously. Two general weaknesses in candidate responses were noted in more than one question part. The first of these was that candidates often ignored the context of a question and gave responses that were either too general or that, whilst true in the general case, did not apply in the specific context of a question. The second was a tendency to attribute human qualities to computers and talk about computers finding tasks easy, or being able to understand something better. Candidates should recognise that computers do not find things easy or understand things, but instead computers and programmers are concerned with aspect of correctness and efficiency such as how quickly an algorithm runs or how straightforward it is for the programmer to code the algorithm. Specific feedback on each question part can be found below. Question 1 Question 1 was about various aspects of computer security. Part (a) was very well answered with the vast majority of candidates correctly identifying that a web server log was a method of accounting and a digital signature was a form of authentication. In part (b), most candidates were able to demonstrate a good understanding of computer viruses, explaining that they were malicious, self-replicating programs that attached themselves to other programs. Understanding of worms was less good, with many candidates failing to recognise that the key feature of a worm is that it distributes itself over a network by exploiting security weaknesses. Many candidates confused worms with Trojans or believed that the distinction between a worm and virus related to the actions that they might take on an infected computer rather than their method of spread. The use of digital signatures, covered in part (c) has been asked about before, and was fairly well understood. Many full mark responses were seen. Even candidates who did not achieve many marks mentioned key terms, suggesting that they had some degree of understanding. Some candidates went on to explain parts of the process that were not asked about on this paper, such as how the entire message might be encrypted or how the digital signature would be authenticated by the recipient. It is important that candidates read the question being asked so that they do not waste time writing material that cannot achieve marks. Question 2 Question 2 was about Reverse Polish Notation (RPN). The vast majority of candidates were able to convert both of the expressions from RPN to infix notation for part (a). Part (b) asked candidates to explain an advantage of RPN. The majority of candidates were able to do this, although some responses were rather superficial. It is true that RPN does not require the use of brackets in expressions, but few candidates went on to explain that this was the case because the order of evaluation was determined entirely by the order in which the operators appeared. Some candidates revealed a lack of understanding of why the lack of the use of brackets was significant, suggesting that it would save memory. Candidates need to be careful to 2 of 8

3 avoid using human-oriented terms such as understand and easier in the context of answers relating to computer programs. Responses such as A computer can understand it more easily were not markworthy. Question 3 Question 3 was about floating point representation of numbers. Parts (a) to (c) which required mathematical working were mostly well answered, but parts (d) and (e) were less well so. The majority of candidates were correctly able to identify the most negative value for part (a). For part (b), candidates had to convert from floating point to decimal. Candidates seemed to find this question part harder than similar question parts in previous years as a result of the magnitude of the mantissa (11). Candidates who worked out the answer using the calculation answer = mantissa * 2exponent appeared to make fewer calculation errors than those who attempted to work out the answer by shifting the binary point by 11 places. Part (c) was very well answered with the vast majority of candidates achieving all three marks. Those who failed to achieve full marks had most commonly arrived at an incorrect representation of 12¾ in fixed point binary at the start of their working or gave the exponent as -4 instead of +4. The advantages of normalising floating point numbers were not well known for part (d) (i). Important advantages include achieving the maximum precision within a given number of bits and facilitating simple/efficient comparison of two numbers because each number has a unique representation. It was not sufficient to state that normalisation offered improved or maximum precision; a candidate had to make clear that this was within a given number of bits. A small number of candidates gave advantages of floating point over fixed point or advantages of normalising a database. For part (d) (ii) candidates were presented with some stimulus material about how floating point representation was used and had to analyse this and then use their understanding of the mantissa and exponent to explain how a more precise representation could be achieved. It was pleasing to see how many candidates correctly identified that one of the two bits was redundant and could be inferred and then went on to explain that the bit that was freed up could be used in the mantissa to store an extra bit of precision. To achieve the second mark, candidates had to make clear that the extra bit would be used in the mantissa, not the exponent, which some failed to do. Some candidates answered a question from a previous paper that had asked about moving a bit between the mantissa and exponent. Such responses were not markworthy on this occasion. Question 4 This question was about data transmission and Mealy machines. For part (a) the vast majority of candidates were able to work out the correct parity bit and give suitable start and stop bits. For part (b), most candidates were able to explain that asynchronous data transmission meant that that the clocks at the sender and receiver were not synchronised, or that there was no common clock. However, the purpose of the start and stop bit were poorly understood. The start bit is used to temporarily synchronise the clock of the receiver to that of the transmitter. The least well understood part of the question was the purpose of the stop bit. The stop bit serves two purposes. The first is to allow the receiver to process the received data, for example, to transfer it out of a 3 of 8

4 receive buffer, before the next transmission is received. The second is to allow the identification of the next start bit, which is why the stop and start bit always have different binary values. Parts (c) and (d) were well answered, with most candidates being able to use the Mealy machine and then recognise the significance of its output. For part (e), the advantages of serial communication were not well understood, with many candidates failing to achieve any marks and only a few achieving both. When answering this type of question, candidates need to be aware of the context given in the question. Good responses recognised that a small amount of data was being sent, so data transmission speed was not a significant factor in this system, and went on to discuss problems associated with parallel communication that would not occur if serial communication was used, notably data skew and crosstalk. Marks were awarded for points relating to costs if these were justified, but really the fact that fewer wires were required so this would be cheaper was a very weak point given that the question stated that the two communicating devices were next to each other. Candidates need to ensure that they consider their response in relation to the context of a question. Question 5 This question was about data structures, with much of the emphasis placed on binary trees. Parts (a) and (b), which related to searching a binary search tree, were both well answered with two thirds of candidates correctly identifying the items that would be examined during the search and over half correctly identifying the time complexity of the search operation for part (b). For part (c) candidates had to represent a binary tree using an array of records. This was well tackled with candidates correctly using pointers to indicate the relationships between the data items. It was not enough to represent leaf node branches with a blank space or a dash for the left and right pointers. An appropriate value such as an unused index number eg 0 or a NULL value was required. Most candidates achieved a reasonable number of marks for question part (d). The key difference between a static and dynamic data structure, that the former had a fixed size that was defined at compile time and that the latter had a variable size which could change at run time was well understood. However, not many candidates went on to explain any other differences, such as the fact that memory space might be wasted if a static structure was relatively empty or that a static structure would generally be allocated consecutive memory locations. The purpose of the heap was well understood, as being a pool of available unused memory that could be allocated to a dynamic structure at runtime. The most common misunderstanding was that the heap was where a dynamic structure stored its data. Some candidates made did not get the mark for explaining the purpose of the heap as whilst they made clear that the heap was used for dynamically allocating memory, their responses did not make clear that the heap was the unused memory rather than the memory that new data was stored in. Part (e) (i) was well answered with the majority of candidates correctly identifying the order that the items would be output. Virtually everyone who correctly identified the order explained the significance of this for part (e) (ii). 4 of 8

5 For part (f) candidates had to explain why graph traversal was a more complex problem than tree traversal. Many responses recognised that features such as cycles and weighting were the key factors that might contribute to this. Some candidates went beyond what was required and provided excellent explanations of, for example, how a cycle in a graph might cause a problem for a traversal algorithm. Question 6 This question was about methods of using software, including rich (thick) and thin client systems and the use of Software as a Service. Most candidates achieved some marks and were able to describe the fundamental difference between the two, ie that rich client systems did processing locally whilst thin client systems relied on processing data on a server. Descriptions of the different hardware requirements of the two were less good. As was the case when a similar question was asked previously, many candidates just referred to better hardware, more powerful machines or even more hardware, none of which were markworthy. To achieve marks candidates needed to make specific points such as a rich client needing to have sufficient secondary storage for all the software. Many responses also only concentrated on the hardware requirements of the workstations, ignoring the server and network which would have been required for a thin client system. With regard to Software as a Service, most candidates knew what this was, ie software that was accessed and run on a remote server, and many candidates wrote a little about it, but few went on to describe many other features of it, such as the fact that it might be provided on a subscription basis rather than as a one-off purchase or that the server would be operated by a separate company who would take care of tasks such as updating the software. A small number of candidates got rich and thin client systems the wrong way around or compared peer-to-peer with server based systems instead of answering the question asked. Another occasionally seen mistake was to believe that thin clients connected to rich clients so the rich clients could do the processing. Question 7 This question was about Turing machines. After being introduced to the specification as a new topic in 2009 this topic is now fairly well understood. For part (a) candidates had to consider which aspects of a modern computer system most closely mapped to the purpose of the transition function and tape in the Turing machine, and for part (b) candidates had to explain the relationship between a Turing machine and an algorithm. About a quarter of candidates achieved both marks for (a). For (b), candidates were expected to recognise that if an algorithm existed for a problem then a Turing machine could be created to solve the problem, and vice-versa. It was not enough to just state that a Turing machine could carry out an algorithm; to be markworthy a response had to make clear that a Turing machine could carry out any algorithm. The trace for part (c) was very well completed with the overwhelming majority of candidates achieving full marks. For part (d) candidates had to explain the purpose of the Turing machine. From just examining the relationship between the input and output, a number of explanations were 5 of 8

6 plausible, but having carried through the process of a trace and having examined the transition function, candidates needed to identify that the Turing machine would reverse the order of the string on the tape. Responses that referred to flipping needed to make clear that it was the order that changed rather than that bits were flipped from 0s to 1s. Question 8 This question was about relational databases. For part (a) candidates had to draw relationships onto an Entity-Relationship diagram. Most candidates achieved at least one mark, but only about a third achieved both. A surprising number drew only two relationships when the question had stated that three were required. For part (b) candidates had to complete an update query. This was well answered. The most common errors were to include quotation marks where they were not required or to be confused about the syntax of the first line, for example by including the fieldname instead of just the table name after the keyword UPDATE. Part (c) required candidates to write a query to retrieve data from the database. The vast majority of candidates understood the basic structure of a query and were able to achieve some marks. Common mistakes were to include the Customer table in the FROM clause, which was not necessary, to miss out the part of the link between the Parcel and PriceBand tables that related to the parcel weight and to put the list into descending instead of ascending order. Some responses were seen that included (ASC) in brackets in the ORDER BY clause. The inclusion of brackets would prevent this from working; the brackets are included in mark schemes to indicate to examiners that the ASC is optional, they are not part of the language syntax. Candidates needed to redesign part of the database to eliminate some redundancy that was identified for part (d). The majority of candidates recognised that a new relation was required to represent the postcodes, but the number who correctly designed the new relation and redesigned the Parcel relation was relatively small. The most common error was to include the house number in the new PostcodeLookup relation instead of leaving it in the Parcel relation. Question 9 This question was about operating systems. Two thirds of candidates were able to identify three appropriate types of management that would be carried out by an operating system for part (a). For part (b), candidates had to explain what an interactive operating system was. To achieve the mark for this question part a candidate had to explain that the OS would be in direct two-way communication with the user. It was not enough to just state that the OS would respond to user inputs, as this would be the case with most types of operating system. Question 10 This question was about abstraction, object-oriented programming and linked lists. For part (a) candidates had to explain how the LinkedList class was a form of abstraction. Many gave a definition of abstraction but failed to apply this to the LinkedList class and so did not achieve a mark. Good responses made clear that the LinkedList class was an example of 6 of 8

7 abstraction because it allowed a programmer to manipulate items in a linked list without having to be concerned about how the linked list was implemented. For part (b) candidates had to explain why the functions and procedures in the class were public whilst the data items were not. Many candidates were able to obtain a mark for the former, but few did so for the latter. Good responses made clear that the functions and procedures were public as they would need to be called from outside of the class to implement the game, and the data items were private so that their values could only be modified in a controlled way from outside of the class, by calling the procedures of the class. It was not sufficient to state that the data items were private because they were only used by the class or because they should not be changed. Candidates had to write an algorithm for deleting an item from a linked list for part (c). A question was asked in a previous year about inserting an item into a linked list and the standard of responses to this question was notably better than was the case in the previous year. The majority of candidates had at least a good attempt at writing the part of the algorithm that would find the correct item to delete and many were then able to change the pointers to delete the item. Common mistakes and omissions were to fail to keep track of the pointer to the previous item when searching, to release the item to delete back to the heap before changing the pointer around it or to increase the current pointer by the fixed value of 1 on each iteration of a search loop. Few candidates scored all eight marks. If a candidate achieved seven but not eight marks this was usually because the algorithm did not take account of the fact that the item to delete might be the first item in the list, in which case the start pointer would need to be changed. Question 11 This question was about the use of hashing. In part (a) candidates had to compare the efficiency of searching a hash table with searching an unordered list. There were many good responses to this which explained that a slow linear search would be required for an unordered list but a fast calculation of a hash value is all that would be needed for the hash table implementation, and using this the location of the translation could be directly found. For part (b) candidates had to explain what a collision was and how it could be dealt with. The majority of candidates appeared to understand both of these but some failed to achieve marks by not stating points explicitly. For example, too many candidates failed to explain the basic point that if two items hashed to the same value then they would be stored at the same location, and the second value would overwrite the first. Various sensible methods of dealing with a collision were well described. Part (c) required candidates to explain why the English word had to be stored in addition to the French word. Some correctly identified that when performing English to French translation, if two English words had hashed to the same value, it would not be possible to tell which the correct translation was unless the English word was stored. A small number of candidates incorrectly believed that the translation was being done in reverse (French to English) and explained that the hash function would be one-way, which whilst true was not a correct answer to the question that had been asked. 7 of 8

8 Question 12 This question was about different formal methods of defining a language. For part (a) some candidates correctly identified that a syntax diagram had been used, but many incorrectly identified the notation as Bakus-Naur Form (BNF). Three quarters of candidates identified that language definition 3 was different to the other two for part (b) and a pleasing number were able to explain that for definition 3 the sign before the binary number was optional in response to part (c). Some candidates gave an almost correct answer to part (c), recognising that the difference related to the sign, despite having identified the wrong language definition in part (b). For part (d) candidates had to write a regular expression to recognise the same language as the finite statue automaton on the question paper. This question part was quite difficult and was poorly tackled. The most common mistake was to assume that any repetitions of ba always had to be made before any repetitions of c and to give an answer like a(ba)*c*. Mark Ranges and Award of Grades Grade boundaries and cumulative percentage grades are available on the Results Statistics page of the AQA Website. Converting Marks into UMS marks Convert raw marks into Uniform Mark Scale (UMS) marks by using the link below. UMS conversion calculator 8 of 8