Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Mark Scheme Method

OCR H446/02 Paper 2 Written Q&A;
Algorithms and programming - written theory only
Compiled from uploaded OCR H446/02 question papers and mark schemes: 2017,
2018, 2019, 2020, 2021, 2022, 2023, 2024 and 2025.

Only Component 2 / Paper 2. No Paper 1 content. No code-writing tasks. No trace

tables. No calculations.

Page 1
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

How to use this PDF

Each answer is written as compact exam-style mark points. The [1] markers show the separate creditworthy
ideas that can earn marks.
Practise by covering the answer, writing a response from memory, then checking whether each [1] point is
present.
For 6-12 mark evaluation questions, learn the answer plan rather than memorising a full essay. OCR rewards
accurate theory, application and a justified judgement.
This booklet includes only H446/02 Algorithms and programming written-theory questions. It deliberately
excludes code completion, class/constructor code, trace tables, practical sorting/searching walkthroughs,
calculations and table-completion tasks.

Coverage
2017 - H446/02 Paper 2, 2018 - H446/02 Paper 2, 2019 - H446/02 Paper 2, 2020 - H446/02 Paper 2, 2021 -
H446/02 Paper 2, 2022 - H446/02 Paper 2, 2023 - H446/02 Paper 2, 2024 - H446/02 Paper 2, 2025 - H446/02
Paper 2

Page 2
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Year-by-year written Q&A;

2017 - H446/02 Paper 2
Question:
Using Big-O notation, state and explain the best-case time complexity of insertion sort.

Answer:
The best-case time complexity is O(n). [1] This occurs when the list is already sorted. [1] As the number of items
increases, the number of comparisons or steps increases linearly. [1] Doubling the number of items would roughly
double the work rather than square it. [1]

Question:
Outline how bubble sort works and explain how its complexity compares with insertion sort.

Answer:
Bubble sort compares adjacent items. [1] If two adjacent items are out of order, they are swapped. [1] It repeats
passes through the list until a full pass makes no swaps. [1] O(n^2) time means the running time grows
quadratically as the data size increases. [1] O(1) space means the amount of extra memory stays constant
regardless of the number of items. [1] Both bubble sort and insertion sort are in-place, so both can have O(1)
space complexity. [1] The same Big-O complexity does not mean the algorithms take exactly the same time. [1]
Insertion sort is usually faster than bubble sort, although neither scales well for very large data sets. [1]

Question:
Explain why a linked list can be suitable for an ordering system.

Answer:
Orders can be processed in the order required by the list. [1] New orders can be inserted at any position, such as
placing a high-priority order earlier. [1] Completed orders can be removed from any position in the list. [1] The list
is dynamic, so orders can be added or deleted as needed. [1]

Question:
Justify the choice of a linear search rather than a binary search for an unordered linked list.

Answer:
The appropriate algorithm is linear search. [1] Linear search does not require the items to be stored in a particular
order. [1] Binary search would not be suitable unless the items were ordered. [1]

Question:
Identify features of an IDE and describe how each benefits the programmer.

Answer:
Auto-complete suggests identifiers and reduces spelling errors. [1] Syntax highlighting colour-codes code so
structures and possible syntax errors are easier to spot. [1] Step mode runs one statement at a time so the
programmer can check the effect of each statement. [1] Breakpoints stop the program at chosen lines so variable
values can be inspected. [1] A watch window shows how variables change while the program runs. [1] Error
diagnostics locate errors and give details to help correction. [1]

Page 3
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Discuss how concurrent programming could be applied to an order-processing system.

Answer:
Concurrent programming allows processes or threads to overlap in time. [1] On a single-core processor it
simulates simultaneous processing by switching between threads. [1] Several tasks may be entered or updated at
overlapping times. [1] Shared data may be accessed by more than one thread. [1] Locking or mutual exclusion is
needed to stop two threads corrupting the same data. [1] This makes the program more complex to write and
debug. [1] Concurrency can model real systems where independent activities occur at the same time. [1] Not all
parts can be parallelised, so extra processors do not produce a proportional speed-up. [1]

Question:
Describe advantages of splitting a program into sub-procedures.

Answer:
Sub-procedures can be reused, so code does not need to be rewritten. [1] Different programmers can work on
different procedures. [1] This can speed up development because work can happen concurrently. [1] Each
procedure can be tested separately, making debugging easier. [1]

Question:
Explain the difference between depth-first traversal and breadth-first traversal.

Answer:
Depth-first traversal follows a branch as far as possible before moving to another branch. [1] In post-order
traversal, a node is visited after its child nodes have been processed. [1] When no further child node can be
followed, the traversal backtracks to a previous node. [1] Breadth-first traversal visits all nodes at the current level
or distance before moving to the next level. [1] Depth-first traversal can use a stack, while breadth-first traversal
uses a queue. [1]

Question:
Explain how backtracking is used in a depth-first traversal.

Answer:
When a node has no unvisited child nodes, the algorithm returns to the previous node. [1] It then checks whether
that previous node has another branch to visit. [1] This repeats until a new unvisited node is found or all reachable
nodes have been visited. [1]

Question:
Describe the process of decomposition.

Answer:
Decomposition is splitting a problem into smaller component parts or sub-problems. [1] Each part can then be
considered, designed, implemented or tested separately. [1]

Page 4
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Discuss the need for and purpose of abstraction in a program such as a virtual pet game.

Answer:
Abstraction removes unnecessary details from the real-world problem. [1] It keeps only the features needed for
the program's purpose. [1] Symbols or simplified values can represent real-world details. [1] Irrelevant details can
be omitted. [1] This reduces programming complexity. [1] It can reduce processing and memory requirements. [1]
It helps the developer focus on the core functionality. [1] Too much abstraction may make the model unrealistic or
less useful. [1]

Page 5
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

2018 - H446/02 Paper 2

Question:
State what is meant by problem recognition and decomposition, and give another computational method.

Answer:
Problem recognition is identifying that there is a problem to solve, or identifying what the problem is. [1]
Decomposition is splitting the problem into smaller sub-problems. [1] Other computational methods include
abstraction, divide and conquer, modelling, heuristics, visualisation, backtracking and concurrency. [1]

Question:
Define data mining and describe how a business could use it.

Answer:
Data mining is finding patterns or useful information in large quantities of data. [1] It can identify customer trends.
[1] The business can use those trends to choose products or offers for customers. [1] It can identify poor-selling
items. [1] The business can replace poor-selling items or change its promotion strategy. [1]

Question:
Define performance modelling and describe one way it can be used.

Answer:
Performance modelling is simulating or testing the behaviour of a system before it is used. [1] It can be used to
test the system with many simultaneous users, orders or data items. [1] This helps identify performance issues
before installation. [1]

Question:
Describe one benefit of creating reusable program components.

Answer:
Reusable components can be used again in future programs. [1] They do not need to be rewritten, saving
development time. [1] If they have already been tested, they can reduce later testing and debugging effort. [1]

Question:
Identify differences between a graph and a tree.

Answer:
A graph may contain cycles, while a tree does not. [1] A tree normally has a hierarchy with a root and parent-child
relationships. [1] A graph may be directed or undirected and may also be weighted. [1]

Question:
Explain how a graph can be an abstraction of a problem.

Answer:
The graph removes the full details of the real problem and represents only the relevant states or stages. [1]
Nodes can represent possible states or sub-problems. [1] Edges can represent possible routes or transitions
between those states. [1]

Page 6
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Identify advantages of using a visualisation to represent a problem.

Answer:
A visualisation presents information in a simpler form for humans to understand. [1] It can make complex
relationships easier to see. [1] It can help plan or explain a solution before implementing an algorithm. [1]

Question:
Compare Dijkstra's algorithm and the A* search algorithm when finding a shortest path.

Answer:
Dijkstra's algorithm uses known edge weights or distances from the start node. [1] A* also uses a heuristic
estimate of the distance to the goal. [1] A useful heuristic can guide A* toward promising paths faster. [1] A* may
need fewer comparisons because it does not have to explore as many alternatives. [1] Dijkstra will find the
shortest path without needing a heuristic. [1] A* depends on the quality of the heuristic, so a poor heuristic
reduces its advantage. [1] A* can be preferable when the graph is large and a reliable heuristic is available. [1]

Question:
Explain why a parameter may be passed by value rather than by reference in a recursive function.

Answer:
Passing by value sends a copy of the value into the sub-program. [1] The original variable is not overwritten by
changes inside the sub-program. [1] Passing by reference gives access to the original variable's memory location.
[1] In recursive functions, passing by reference could cause values from earlier calls to be overwritten and
produce incorrect results. [1]

Question:
Compare the use of parameters and global variables in recursive functions.

Answer:
Parameters allow values to be passed into a sub-program. [1] A parameter becomes local to that call of the
function. [1] Global variables can be accessed throughout the program. [1] Using global variables in recursion
risks each call overwriting values needed by other calls. [1] Parameters make recursive calls independent
because each call has its own local values. [1] Global variables occupy memory for the whole runtime of the
program. [1] Parameters or local variables can be released when the call finishes. [1]

Question:
Explain why a recursive algorithm may use more memory than an iterative algorithm.

Answer:
Each recursive call stores its current state on the call stack. [1] This may include parameters, local variables and
the return address. [1] Iteration normally reuses the same variables each time the loop repeats. [1]

Question:
Describe the decisions needed before pushing to or popping from a stack.

Answer:
Before push, the program checks whether the stack is full. [1] If it is not full, the new item can be inserted. [1] If it
is full, an error should be reported or the push should be rejected. [1] Before pop, the program checks whether the
stack is empty. [1] If it is not empty, the top item can be returned or removed. [1] If it is empty, an error should be
reported or the pop should be rejected. [1]

Page 7
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Describe how a 1D array can be used to implement a stack.

Answer:
The array size is defined. [1] A stack pointer stores the position of the top item. [1] When an item is pushed, it is
stored at the top position and the pointer is updated. [1] When an item is popped, the top item is returned or
removed and the pointer is updated in the opposite direction. [1]

Question:
Describe the stages of a binary search on a sorted array.

Answer:
Set lower and upper bounds for the search range. [1] Calculate the midpoint from the bounds. [1] Compare the
middle item with the target value. [1] If the middle item equals the target, return the position or set found to true.
[1] If the middle item is less than the target, move the lower bound above the midpoint. [1] If the middle item is
greater than the target, move the upper bound below the midpoint. [1] Repeat until the item is found or the bounds
cross. [1]

Question:
Explain similarities and differences between a record and a class.

Answer:
A record is a data structure that stores related fields. [1] A class is a template for creating objects. [1] Both can
store data of different types, accessed using field or attribute names. [1] A class can also include methods that
define behaviour. [1] A class can use visibility such as private attributes, whereas a record usually just groups
data. [1]

Question:
Define the term queue.

Answer:
A queue is a data structure. [1] It is first in, first out, so the first item added is the first item removed. [1]

Question:
Explain how private attributes improve data integrity.

Answer:
Private attributes are encapsulated and can only be accessed through methods. [1] The methods can validate
data before changing the attribute. [1] This reduces accidental or inappropriate changes to the data. [1]

Question:
Describe how a program can make sure data still exists the next time it is run.

Answer:
Save the data to an external file when the program closes. [1] Load the saved data from the file when the program
starts again. [1]

Page 8
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Evaluate the use of caching and concurrent processing for a very large searchable data set.

Answer:
Caching stores previously used data in a faster location such as RAM. [1] If the same data is needed again, it can
be retrieved more quickly. [1] Caching is most useful when the same items are searched or accessed repeatedly.
[1] With very large data sets, caching everything may be unrealistic. [1] Concurrent processing allows several
processes or processors to work at overlapping times. [1] A large linear search could be split so different
processors search different sections. [1] This may reduce search time, but the speed-up is limited by bottlenecks
such as storage access. [1] Use caching for repeated searches and concurrency for large independent searches.
[1]

Page 9
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

2019 - H446/02 Paper 2

Question:
Explain why a queue is used instead of a stack when data must be accessed in arrival order.

Answer:
A queue is first in first out. [1] It retrieves values in the same order they were recorded. [1] A stack is last in first
out. [1] It would retrieve the values in reverse order, which is unsuitable for arrival-order processing. [1]

Question:
Explain why dequeue() is a function rather than a procedure.

Answer:
dequeue() is a function because it returns the first item from the queue. [1] A procedure would carry out an action
without returning a value to the calling code. [1]

Question:
Describe what O(n), O(n^2) and O(1) mean for bubble sort.

Answer:
O(n) is linear time. [1] In the best case, the time grows in proportion to the number of items. [1] This occurs when
the data is already in order. [1] O(n^2) is quadratic time. [1] In the average or worst case, the time grows in
proportion to the square of the number of items. [1] Worst case is usually when the data begins in reverse order.
[1] O(1) space is constant extra memory. [1] Bubble sort uses the same amount of additional memory regardless
of the number of items. [1]

Question:
Explain the difference between branching and iteration.

Answer:
Branching chooses which code path is run, such as an if statement. [1] The selected branch runs once unless
another control structure repeats it. [1] Iteration repeatedly runs the same code while, until or for a condition. [1]

Question:
Describe the arithmetic operation MOD.

Answer:
MOD gives the remainder after integer division. [1] For example, 10 MOD 3 is 1 because 3 fits into 10 three times
with remainder 1. [1]

Question:
State one benefit and one drawback of using iteration instead of recursion.

Answer:
Iteration may run faster because it avoids repeated function calls. [1] Iteration cannot run out of call-stack space in
the way deep recursion can. [1] However, iteration can make code longer or more complex. [1] Some problems
are expressed more clearly or elegantly using recursion. [1]

Page 10
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Explain how decomposition can aid the design of a game program.

Answer:
Decomposition splits the game into sub-problems. [1] This makes the design more manageable and easier to
understand. [1] Each sub-problem can be tackled independently. [1]

Question:
Define pipelining and give an example of its use in a program.

Answer:
Pipelining is where the result from one process or procedure is passed into the next process. [1] For example, an
event-detection procedure can feed into another procedure that updates the program state. [1]

Question:
Discuss the use of an IDE rather than only a text editor when developing a program.

Answer:
An IDE provides tools that help write code, such as syntax colouring, predictive text and auto-correct. [1] These
can reduce spelling or syntax errors. [1] It provides debugging tools such as stepping, breakpoints and variable
watch windows. [1] These help locate logic errors by inspecting the program during execution. [1] The program
can often be run without leaving the development environment. [1] An IDE may integrate tools such as version
control for team development. [1] A text editor may be simpler and use fewer resources, but provides less
support. [1] An IDE is usually more appropriate for complex program development because debugging and
project support outweigh the overhead. [1]

Question:
Explain why insertion sort might use less memory than merge sort.

Answer:
Merge sort may create new arrays when it splits and merges the data. [1] Merge sort is often implemented
recursively, which adds data to the call stack. [1] Insertion sort sorts in place, so it does not require additional
arrays. [1]

Question:
Evaluate the use of data mining to improve a social networking website.

Answer:
Data mining extracts useful information from large data sets. [1] It looks for patterns or specific occurrences in
stored data. [1] It can identify which features users use most. [1] It can identify features that are rarely used and
may need redesign or removal. [1] It can identify characteristics of user groups to inform new features. [1] This
may help target advertising or improve user engagement. [1] It can save time and money by focusing
development on popular areas. [1] It must be handled carefully because privacy law and user trust are significant
issues. [1] Data mining is useful if there is enough relevant data and it is used ethically and legally. [1]

Page 11
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Give examples of abstraction in a program interface and explain why abstraction is needed.

Answer:
Abstraction can replace real images with labelled shapes. [1] It can show a simplified layout rather than every
real-world detail. [1] It removes irrelevant features that are not needed by the program. [1] This reduces the
complexity of the design. [1] It can reduce programming effort. [1] It can reduce memory and processing
requirements because detailed images or unnecessary information are not stored. [1]

Question:
Explain and evaluate the use of caching and reusable components in a program design.

Answer:
Caching stores data that has been used in a fast-access location in case it is needed again. [1] This can speed up
repeated access to items, requirements or layouts. [1] Reusable components are pieces of code that can be
called many times or used in multiple places. [1] Subroutines, classes and libraries can be reusable components.
[1] Reusable components can make later adaptation easier. [1] Caching can improve performance but uses extra
memory. [1] Reusable components require more careful upfront design. [1] Both are worthwhile when repeated
operations or repeated item types are expected. [1]

Page 12
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

2020 - H446/02 Paper 2

Question:
State what is meant by abstraction and describe how it can be used in a tree model.

Answer:
Abstraction removes unnecessary detail from the problem. [1] A tree can represent real movements or states
using nodes and edges. [1] This hides details about the full real-world move and keeps only the information
needed by the algorithm. [1]

Question:
State why a tree may not be a binary search tree.

Answer:
A binary tree must have no more than two child nodes from any node. [1] A binary search tree must also be
ordered so values less than a node are on one side and greater values are on the other side. [1]

Question:
State the type of pointer used when a node does not point to any other node.

Answer:
A null pointer is used when a node has no next child or next node. [1]

Question:
Give similarities and differences between a tree and a graph.

Answer:
Both trees and graphs consist of nodes. [1] Both use edges or links to connect nodes. [1] A tree has a root node,
whereas a graph does not need one. [1] A tree does not contain cycles, whereas a graph can contain cycles. [1]

Question:
Explain why decomposition can help during program development.

Answer:
Decomposition splits a problem into smaller sub-problems. [1] Repeated decomposition creates smaller parts that
are easier to solve. [1] These parts can become modules or subroutines in the program. [1] Work can be divided
between programmers or teams so development can happen concurrently. [1]

Question:
Discuss the need for concurrent processing in an online ticket-selling system.

Answer:
Concurrent processing allows several processes or threads to be handled at overlapping times. [1] Multiple
customers may send requests to the server at the same time. [1] The server must allow several users to browse,
reserve or purchase tickets without waiting for one user to finish completely. [1] Record locking may be needed so
two customers cannot buy the same seat. [1] This improves responsiveness and supports realistic online
demand. [1] However, it makes programming more complex because shared data must be controlled safely. [1]

Page 13
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Compare recursion and iteration.

Answer:
Recursion solves a problem by a routine calling itself. [1] Iteration repeats a block of code using a loop. [1]
Recursion uses stack memory because each call stores its own state. [1] Iteration usually reuses the same
variables and may use less memory. [1] Recursion can express some problems more elegantly, but it can run out
of stack space if too many calls are made. [1]

Question:
Explain why an array parameter may be passed by reference rather than by value.

Answer:
Passing by reference allows the subroutine to change the original array. [1] When control returns to the main
program, the changes are still present. [1] Passing by value would create a copy, so the original array would not
be changed. [1] Passing by reference also avoids copying the whole array, which can use less memory. [1]

Question:
Compare merge sort, quick sort and insertion sort for small and large arrays.

Answer:
Insertion sort has simple logic and can be efficient for a small or nearly sorted array. [1] Its average and
worst-case time complexity is O(n^2), so it does not scale well for large arrays. [1] Merge sort has O(n log n) time
complexity, so it is more suitable for large arrays. [1] Merge sort may require extra memory because it creates or
stores sublists. [1] Quick sort has good average time complexity of O(n log n), but its worst case can be O(n^2).
[1] For large unsorted arrays, merge sort or quick sort is usually more suitable than insertion sort. [1]

Question:
Describe how a bubble sort works.

Answer:
A bubble sort compares adjacent pairs of items. [1] If a pair is in the wrong order, the two items are swapped. [1]
This continues to the end of the list for one pass. [1] After a pass, one item has moved to its correct position. [1]
The algorithm repeats passes until no swaps are needed or the maximum number of passes has been reached.
[1]

Question:
State the purpose of the head and tail pointers in a queue.

Answer:
The head pointer identifies the first item in the queue or the next item to remove. [1] The tail pointer identifies the
last item in the queue or where the next item will be added. [1]

Question:
Describe how a queue can be changed to work as a circular queue.

Answer:
When the head or tail pointer moves beyond the final array index, it is reset to 0. [1] This allows free spaces at the
start of the array to be reused. [1] A full queue check must detect when the tail would catch up with the head. [1]

Page 14
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Describe how a program could deal with different priorities in a queue.

Answer:
The program could use separate queues for different priority levels. [1] New items would be added to the queue
matching their priority. [1] Items from the highest-priority queue would be processed before lower-priority items.
[1] Alternatively, a linked list could insert higher-priority items closer to the front by adjusting pointers. [1]

Question:
State the purpose of the free list pointer and head pointer in a linked list.

Answer:
The free list pointer points to the next free node that can be used when adding data. [1] The head pointer points to
the first node in the linked list. [1]

Question:
Describe the difference between get methods and set methods.

Answer:
A get method returns or gives access to an attribute value. [1] A set method changes an attribute value, usually
using a parameter. [1]

Question:
Describe how a linked list can be searched by following pointers.

Answer:
The search starts at the head pointer. [1] The data in the current node is compared with the item being searched
for. [1] If it is not found, the next pointer is followed to the next node. [1] This repeats until the item is found or a
null pointer is reached. [1] A suitable found or not-found message can then be returned. [1]

Question:
Describe IDE features that help with debugging.

Answer:
Stepping runs the program one line at a time so the programmer can locate where an error occurs. [1]
Breakpoints stop execution at chosen lines so a section of code can be inspected. [1] A variable watch window
shows variable values as the program runs. [1] Syntax error highlighting or diagnostics helps identify syntax
errors in the code. [1]

Question:
Discuss how object-oriented techniques can make a linked list reusable.

Answer:
A class is a template that defines the attributes and methods for a linked list or node. [1] Objects can be
instantiated from the class whenever a new list or node is needed. [1] Encapsulation protects attributes by making
them private and controlling access through methods. [1] Inheritance allows a subclass to reuse attributes and
methods from a base class. [1] Overriding allows a subclass to replace a parent method with a more specific
version. [1] These techniques improve reuse, maintenance and modification of a linked-list library. [1]

Page 15
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

2021 - H446/02 Paper 2

Question:
Describe how abstraction can be used when designing a game.

Answer:
Abstraction removes unnecessary details from the real-world situation. [1] Unneeded character or environment
features can be left out. [1] Complex real-world objects can be simplified into specific objects, shapes or stored
attributes. [1] This makes the game easier to design and program. [1]

Question:
Describe what is meant by concurrent processing.

Answer:
Concurrent processing means multiple processes appear to execute at the same time. [1] This can be done by
giving each process a slice of processor time or by using multiple processors. [1]

Question:
Explain why concurrent processing is needed in an online multiplayer system.

Answer:
Many users may send requests to the server at the same time. [1] The server must respond in a reasonable time
while still handling other users. [1] Concurrent processing allows different requests to be handled without one
user blocking all others. [1] Shared data may need protection, such as record locking, so users do not overwrite
each other's changes. [1]

Question:
Define a heuristic in relation to the A* algorithm.

Answer:
A heuristic is an estimate or rule of thumb. [1] In A*, it estimates the remaining distance or cost from a node to the
destination. [1] It helps decide which path to follow first, making the search more efficient. [1]

Question:
Evaluate how data mining can be used to improve a large online game.

Answer:
Data mining searches large amounts of data for useful patterns, trends or relationships. [1] It can identify how
users behave, such as when they play, which features they use and which actions they perform most often. [1]
This can guide future changes, such as adding popular features or removing unused ones. [1] It may increase
user engagement and revenue by making the game more appealing. [1] However, there may be privacy concerns
if user activity is collected or misused. [1]

Question:
State suitable Big O complexities for binary and linear search.

Answer:
A linear search has O(n) average time because the number of comparisons grows linearly with the number of
items. [1] A binary search has O(log n) average time because the search space is repeatedly halved. [1] Both
linear and binary search can have O(1) space complexity when no extra data structure is created. [1]

Page 16
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Identify when a linear search is more appropriate than a binary search.

Answer:
A linear search is more appropriate when the data is not sorted. [1] It can also be suitable for a very small list
where the overhead of binary search is not worthwhile. [1]

Question:
Describe how quicksort sorts data into ascending order.

Answer:
Quicksort chooses a pivot value. [1] Each value is compared with the pivot. [1] Values less than the pivot are
placed in the left sublist. [1] Values greater than the pivot are placed in the right sublist. [1] The same process is
repeated on each sublist until the data is sorted. [1]

Question:
Explain why quicksort is a divide-and-conquer algorithm.

Answer:
Quicksort divides the data into smaller subsets. [1] Each subset is sorted separately. [1] The sorted subsets are
then combined to form the final sorted list. [1]

Question:
Discuss the benefits of using an IDE rather than a text editor.

Answer:
An IDE combines tools such as an editor, compiler and run-time environment in one program. [1] Syntax
highlighting and error diagnostics help identify mistakes more quickly. [1] Auto-complete or auto-correct can
reduce spelling and syntax errors. [1] Breakpoints allow execution to stop at a chosen line during testing. [1]
Stepping lets the programmer run one line at a time to trace logic. [1] A variable watch window shows variable
values while the program runs. [1] This can reduce development and debugging time compared with using a plain
text editor. [1]

Question:
State the purpose of a constructor.

Answer:
A constructor creates or initialises a new instance of a class. [1]

Question:
Describe inheritance using a base class and subclasses.

Answer:
Inheritance occurs when a child or derived class takes attributes and methods from a parent or base class. [1]
The child class can also add extra attributes or methods of its own. [1]

Page 17
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Explain why pipelining can improve processor performance.

Answer:
Pipelining allows one instruction to be fetched while another is decoded and another is executed. [1] This reduces
the time that parts of the processor are idle. [1] More instructions can be processed over the same period of time.
[1]

Question:
State the purpose of two pointers used in a circular queue.

Answer:
One pointer identifies the first item in the queue, which is the next item to be removed. [1] Another pointer
identifies the last item or next free position where a new item will be added. [1]

Question:
Describe how enqueue works in a circular queue.

Answer:
The algorithm checks whether the queue is full before adding data. [1] If the queue is full, it returns a suitable
failure value. [1] If there is space, the new item is stored at the tail or next free position. [1] The tail pointer is then
moved to the next position. [1] If the tail moves beyond the last array index, it wraps around to 0. [1]

Question:
Explain why decomposing a problem can help a developer design a solution.

Answer:
Decomposition splits the problem into smaller chunks. [1] Smaller problems are more manageable and easier to
solve. [1] It can show where code can be reused. [1] It can also allow different programmers to work on different
parts. [1]

Question:
Describe the purpose of branching and iteration in an algorithm.

Answer:
Branching allows the program to make decisions based on a condition. [1] It can choose different actions
depending on stored values or user input. [1] Iteration repeats a set of instructions. [1] It is useful when each
element in a data structure must be processed. [1]

Question:
Compare global and local variables, parameters and efficiency.

Answer:
A local variable has scope only within the module or subroutine where it is declared. [1] A global variable can be
accessed throughout the program. [1] Local variables are removed from memory when the module ends, so they
can be more memory efficient. [1] Global variables remain in memory throughout execution. [1] If data is local, it
must be passed as a parameter or returned from a function. [1] Passing by reference gives access to the original
data, while passing by value sends a copy. [1] Global variables can make small programs simpler, but they can
reduce modularity and make errors harder to trace. [1]

Page 18
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Describe how a program could automatically generate a grid-based puzzle.

Answer:
The program can use random numbers to generate the grid size. [1] It can generate a horizontal size and vertical
size for the grid. [1] It can loop through each cell in the grid. [1] For each cell, it can randomly store a value such
as 0 or 1 to represent the cell state. [1]

Page 19
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

2022 - H446/02 Paper 2

Question:
State the precondition for a binary search.

Answer:
The data being searched must be in order or sorted. [1]

Question:
Describe how a linear search works.

Answer:
The search item is compared with the first value in the list. [1] If it is not found, the search item is compared with
the next value. [1] This repeats until the item is found or the end of the list is reached. [1] The algorithm returns
the position if the item is found, or a suitable not-found value if it is not found. [1]

Question:
Explain the difference between passing by value and passing by reference.

Answer:
Passing by value gives the subprogram a copy of the original data. [1] Changes made inside the subprogram
affect only the copy. [1] Passing by reference gives the subprogram access to the original memory location. [1]
Changes made inside the subprogram can alter the original data and remain after the subprogram ends. [1]

Question:
Give one benefit and one drawback of using a global variable.

Answer:
A global variable can be accessed from any subprogram, so it does not need to be passed repeatedly as a
parameter. [1] A global variable remains in memory for the whole program execution, increasing memory use. [1]
It can also be altered unexpectedly by another part of the program, causing side effects. [1]

Question:
Identify IDE features that help programmers write and test code.

Answer:
Auto-complete suggests commands or identifiers, reducing typing and spelling errors. [1] Auto-indent formats
code inside structures, improving readability and reducing structural mistakes. [1] Syntax highlighting
colour-codes keywords and structures so errors are easier to spot. [1] Breakpoints stop the program at a chosen
line so variable values can be inspected. [1] A variable watch window displays values while the program runs. [1]
Stepping runs one line at a time to help locate logic errors. [1]

Question:
Describe benefits of creating reusable algorithms or program components.

Answer:
Reusable components save time because the same algorithm does not need to be written repeatedly. [1] They
reduce testing effort if the component has already been tested. [1] They can be used in other programs or stored
in a program library. [1]

Page 20
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Compare breadth-first and depth-first traversal of a tree.

Answer:
Breadth-first traversal visits the root and then all nodes connected at the next level before moving deeper. [1]
Depth-first traversal follows a branch as far as possible before moving to another branch. [1] Depth-first traversal
often uses backtracking when a leaf node is reached. [1] Breadth-first traversal is often better when the required
item is near the root. [1] Depth-first traversal may be better when the required item is deeper in the tree. [1]
Depth-first traversal can be written recursively, but very large trees may use substantial stack space. [1]

Question:
Define abstraction and decomposition.

Answer:
Abstraction is removing unnecessary detail and focusing only on the parts needed to solve the problem. [1]
Decomposition is breaking a problem down into smaller sub-problems. [1]

Question:
Explain why abstraction is useful when designing a program.

Answer:
Abstraction reduces memory requirements by removing unneeded detail. [1] It reduces processing requirements
because unnecessary features do not need to be handled. [1] It simplifies the problem being solved so the
programmer can focus on the important elements. [1]

Question:
Describe how caching can improve an algorithm or system.

Answer:
Caching stores data that has already been used in a faster location such as RAM or cache. [1] If the same data is
needed again, it can be retrieved more quickly. [1] Caching is useful when the same values, designs or results are
repeatedly accessed. [1]

Question:
Explain the difference between a directed graph and an undirected graph.

Answer:
In a directed graph, an edge or arc can only be followed in the direction shown. [1] In an undirected graph, an
edge can be followed in both directions. [1]

Question:
Give differences between a graph and a tree.

Answer:
A graph can have more than one path between nodes, while a tree has a hierarchical structure. [1] A graph does
not need a root node, while a tree has a root node. [1] A graph can contain cycles or loops, while a tree does not.
[1] A graph can be weighted, whereas a tree is normally used to show parent-child relationships. [1]

Page 21
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Explain how a graph can be an abstraction of a real problem.

Answer:
A graph can use nodes to represent important places, states or data items. [1] Edges can represent possible
connections between those nodes. [1] The graph leaves out unnecessary physical detail, such as the exact shape
of real routes. [1]

Question:
Compare Dijkstra's algorithm and A* as path-finding algorithms.

Answer:
Both algorithms can be used to find a shortest path. [1] Dijkstra's algorithm uses known edge weights from the
start node. [1] A* also uses a heuristic estimate of the distance to the target. [1] A* is usually more efficient when
the heuristic guides the search towards the target. [1] Dijkstra's algorithm does not need a heuristic. [1]

Question:
Define performance modelling and describe how it can be used.

Answer:
Performance modelling simulates or models the behaviour of a system before it is used. [1] It can test how the
system behaves with large or small inputs. [1] It can model how well the system scales as the number of users,
deliveries, routes or data items increases. [1]

Question:
Explain why a function is used rather than a procedure.

Answer:
A function is used when a value needs to be returned to the calling code. [1] A procedure carries out a task but
does not return a value in the same way. [1]

Question:
Identify features of a recursive function.

Answer:
A recursive function calls itself. [1] It has a base case or stopping condition that terminates the recursion. [1] Each
recursive call stores its own state, such as parameters and local variables, on the call stack. [1]

Question:
Describe the purpose of a stack pointer.

Answer:
A stack pointer identifies the top item in the stack. [1] It may also identify the next free space where a new item
will be pushed. [1] The pointer is updated when items are pushed or popped. [1]

Page 22
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Describe how head and tail pointers are used in a queue.

Answer:
The head pointer identifies the item that will be dequeued next. [1] The tail pointer identifies where the next item
will be enqueued. [1] Enqueueing updates the tail pointer. [1] Dequeueing updates the head pointer. [1]

Question:
Compare object-oriented and procedural approaches for implementing data structures.

Answer:
Object-oriented programming defines data structures as objects with attributes and methods. [1] Attributes can be
made private to reduce accidental changes. [1] Multiple instances of the same structure can be created without
rewriting the class. [1] Procedural programming executes statements and subroutines in sequence. [1] Procedural
code may require the structure to be passed into subroutines repeatedly. [1] Object-oriented programming can
reduce duplicated code and improve maintainability. [1]

Page 23
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

2023 - H446/02 Paper 2

Question:
Give characteristics of a tree data structure.

Answer:
A tree is a hierarchical structure. [1] Data is stored in nodes. [1] Nodes are linked by branches or edges. [1] A tree
has a root node. [1] Nodes may have child nodes beneath them. [1] Nodes with no children are leaf nodes. [1] A
tree has no cycles or loops. [1]

Question:
Describe how a leaf node is deleted from a binary search tree.

Answer:
The tree is searched until the required leaf node is found. [1] The parent node's pointer to that leaf node is set to
null. [1] The deleted node can then be added to the free storage list or left for garbage collection. [1]

Question:
Describe how a binary search tree can be searched for a value.

Answer:
The search value is first compared with the root node. [1] If the root equals the search value, the value has been
found. [1] If the search value is less than the current node, the left subtree is searched. [1] If the search value is
greater than the current node, the right subtree is searched. [1] This process repeats until the value is found or
there are no further branches to follow. [1]

Question:
Explain how backtracking is used in a depth-first traversal.

Answer:
When a leaf node is reached, there are no further child nodes to visit from that branch. [1] The traversal returns to
the leaf node's parent. [1] It then backtracks to the most recent node that still has an unvisited child. [1]

Question:
Explain problem recognition and decomposition as computational thinking methods.

Answer:
Problem recognition identifies that there is a problem to solve and determines exactly what the problem is. [1] It
also considers whether the problem can be solved using computational methods. [1] Decomposition splits the
problem into smaller sub-problems that can be solved independently. [1] This helps identify inputs, outputs,
constraints and the main features that the program must handle. [1]

Question:
Explain the purpose of linked-list pointers such as headPointer, freeListPointer and null.

Answer:
The head pointer identifies the first element in the linked list. [1] The free-list pointer identifies the next free index
where new data can be stored. [1] A null pointer shows that a node does not point to another node. [1] Null can
therefore indicate the end of a linked list. [1]

Page 24
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Describe benefits of reusable functions or subroutines.

Answer:
One piece of code can be used many times in different places. [1] The same code does not need to be written
repeatedly. [1] Development takes less time because less code must be planned and coded. [1] Maintenance is
easier because a correction in one reusable routine affects every place it is used. [1]

Question:
Describe features of recursion.

Answer:
A recursive function calls itself. [1] Each recursive call creates a new copy of the function's values. [1] These
values are stored on a stack until the call returns. [1] A base case stops further recursive calls. [1]

Question:
Explain common Big O complexities.

Answer:
Big O describes how time or memory requirements change as the amount of data increases. [1] Constant
complexity does not change as the input size grows. [1] Linear complexity grows in proportion to the input size. [1]
Logarithmic complexity grows at a decreasing rate as the input size grows. [1] Exponential complexity grows very
quickly and can become unmanageable for large inputs. [1]

Question:
Define concurrent processing and give benefits.

Answer:
Concurrent processing allows processes to overlap in time. [1] One process can start before another process has
finished. [1] Processes may be given slices of processor time or run on different cores. [1] It can make better use
of processor time and reduce idle time. [1] It can allow a user to interact with the computer while other tasks
continue. [1]

Question:
Describe how merge sort works.

Answer:
The list is split into two smaller lists. [1] These sublists are repeatedly split until each sublist contains one item. [1]
The first items in two sublists are compared. [1] The smaller item is selected and written to a new list. [1] This
merging process repeats until all sorted sublists are recombined into one sorted list. [1]

Question:
Give one benefit and one drawback of merge sort compared with bubble sort.

Answer:
Merge sort has a more efficient time complexity for large data sets, such as O(n log n) rather than O(n^2). [1]
Merge sort uses divide and conquer and can be suitable for concurrent processing. [1] However, merge sort is
harder to implement than bubble sort. [1] Merge sort usually has worse space complexity because it needs extra
lists or stack space. [1]

Page 25
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Give IDE features and explain how they help debugging.

Answer:
Error diagnostics help locate and fix errors. [1] Breakpoints stop the program at chosen points so values can be
inspected. [1] Syntax highlighting identifies keywords, variables and likely syntax errors. [1] Stepping runs the
program line by line so the effect of each statement can be checked. [1] A variable watch window shows how
variable values change during execution. [1] Auto-complete reduces typing errors by completing commands or
identifiers. [1]

Question:
Define abstraction and explain why it is useful.

Answer:
Abstraction removes unnecessary detail from a problem. [1] It allows programmers to focus on the core aspects
of the problem. [1] It simplifies a complex problem so it is easier to solve. [1] It can reduce programming time. [1] It
can reduce memory use or computational power required. [1]

Question:
Define encapsulation.

Answer:
Encapsulation is the bundling of data with the methods that operate on that data. [1] It allows an attribute to be
accessed or changed only through methods. [1] This protects data from inappropriate direct access. [1]

Question:
Describe benefits of object-oriented programming.

Answer:
Classes can be reused in other programs. [1] Inheritance can extend existing classes instead of rewriting them.
[1] Classes can be modified or extended, improving maintainability. [1] Encapsulation makes debugging easier
because it limits how attributes can be changed. [1] Access to attributes can be restricted, improving security and
data integrity. [1] Classes can be distributed between team members, supporting team development. [1]

Question:
Compare local variables, global variables and parameters.

Answer:
A local variable can only be accessed within the subprogram or main program where it is declared. [1] A global
variable can be accessed by all subprograms. [1] A parameter is an item passed into a subprogram. [1] Passing
by value sends a copy, so the original value is not changed. [1] Passing by reference sends access to the original
value, so the original can be changed. [1] Global variables remain in memory throughout execution, which can
increase memory use. [1] Local variables and parameters reduce unwanted side effects because access is more
controlled. [1]

Page 26
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

2024 - H446/02 Paper 2

Question:
Describe what is meant by a variable.

Answer:
A variable is a named memory location. [1] It stores a value or item of data. [1] The value stored in a variable can
be changed while the program runs. [1]

Question:
Describe the difference between a do loop and a while loop.

Answer:
A while loop checks its condition at the start of the loop. [1] This means the code in a while loop may never run.
[1] A do loop checks its condition at the end of the loop. [1] This means the code in a do loop will run at least
once. [1]

Question:
Evaluate the use of data mining.

Answer:
Data mining analyses patterns and anomalies in large data sets. [1] It turns large quantities of data into useful
information that may not be obvious to a human reader. [1] The information can be used to make predictions,
improve services, target advertising or increase revenue. [1] Data mining can identify which features are used
most or least. [1] It can identify difficult tasks or weak performance areas. [1] However, data mining finds patterns
but does not always explain why they occur. [1] Users may have privacy concerns if their activity is logged. [1]
Large-scale data mining can require significant processing power and security controls. [1] The data collection
must be legal and covered by appropriate terms or consent. [1]

Question:
Give differences between trees and graphs.

Answer:
A tree has one root node, while a graph does not have to have a root node. [1] A tree does not allow cycles or
loops, while a graph can. [1] A tree stores data hierarchically, while a graph does not need a hierarchy. [1] A tree
is connected, while a graph can be connected or disconnected. [1] A tree is normally undirected, while a graph
may be directed. [1]

Question:
Explain what heuristics are and when they are useful.

Answer:
A heuristic is a rule of thumb or educated guess. [1] It reduces the time taken to solve a problem by avoiding
exhaustive search. [1] It can find a solution that is good enough or close to the best solution. [1] In A*, a heuristic
can estimate the distance to the destination. [1] Heuristics are useful for complex, large-scale or time-critical
problems. [1] They may be less accurate because they do not examine every possibility. [1] They are suitable
where a perfect answer is not essential, such as many game AI decisions. [1]

Page 27
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Describe the roles of headPointer and tailPointer in a queue.

Answer:
The head pointer identifies the first item in the queue. [1] It also identifies which item will be dequeued next. [1]
The tail pointer identifies the next free space in the queue. [1] It also identifies where the next item will be
enqueued. [1]

Question:
Explain why a queue may be unsuitable when data needs to be resequenced.

Answer:
A queue is a first-in, first-out structure. [1] Items are processed in the order in which they were entered. [1] New
items can only be added at the rear of the queue. [1] The contents cannot be easily sorted or resequenced
without rewriting or using another structure. [1]

Question:
Describe how a pop operation works on a stack.

Answer:
The algorithm first checks whether the stack is empty. [1] If the stack is empty, it returns a suitable value or
warning. [1] If the stack is not empty, the stack pointer is decremented. [1] The value at the top position is
returned from the stack array. [1]

Question:
Describe how a linear search works.

Answer:
The first element is compared with the search item. [1] If the element matches the search item, the algorithm
returns the index or reports that it has been found. [1] If it does not match, the algorithm moves to the next
element. [1] This repeats until the item is found or the end of the list is reached. [1]

Question:
Explain the Big O terms linear, logarithmic, constant and exponential.

Answer:
Linear time means the time taken increases in direct proportion to the number of items. [1] Logarithmic space
means the additional memory grows at a decreasing rate as the number of items increases. [1] Constant
complexity means the requirement stays the same regardless of input size. [1] Exponential complexity means the
requirement grows at a rapidly increasing rate as input size increases. [1]

Question:
Explain why a record data structure can be suitable for storing related data.

Answer:
A record can store multiple items of data under one identifier. [1] A record can store fields of different data types.
[1] This makes it suitable when one item has several related properties, such as text, integer and real values. [1]

Page 28
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Give characteristics and benefits of a binary search tree.

Answer:
Each node in a binary search tree can have at most two child nodes. [1] Values are ordered so lower values are
placed to the left and higher values to the right. [1] The position where a new node is added depends on its order.
[1] Searching can be faster than searching an unordered structure. [1] The structure does not need to be sorted
again each time a new item is inserted. [1]

Question:
Compare depth-first post-order traversal and breadth-first traversal.

Answer:
Breadth-first traversal visits all nodes at the current depth before moving to the next depth. [1] Depth-first traversal
follows one branch to the end before moving to another branch. [1] Depth-first traversal uses backtracking. [1] In
post-order traversal, the root node is output last. [1] Breadth-first traversal does not output a binary search tree in
ascending order. [1]

Question:
Define abstraction and give benefits of using it.

Answer:
Abstraction is the removal of unnecessary detail. [1] It can simplify the problem, algorithm or program code. [1] It
can make the program faster to create. [1] It can reduce memory or processor use in the final program. [1] It
allows the programmer to focus on the core aspects of the problem. [1]

Question:
Define decomposition and give benefits of using it.

Answer:
Decomposition is splitting a problem into smaller sub-problems. [1] It helps identify individual components of the
solution. [1] It can show which components can be tackled concurrently. [1] It can identify reusable program
elements so the same algorithm is not created twice. [1] It allows work to be split between team members. [1] It
makes each smaller problem easier to write, test and debug. [1]

Question:
Explain why an array can be suitable for storing a fixed sequence of related objects.

Answer:
An array has a fixed number of positions when the number of required values is known. [1] It stores data linearly,
matching data that is arranged in a sequence. [1] Array elements can be directly accessed by index. [1] The array
can be iterated through to apply an operation to each position. [1] Its contents are mutable, so elements can be
added, changed or removed. [1]

Page 29
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Evaluate using global variables rather than local variables and parameters.

Answer:
Global variables are created when the program starts and can be accessed or updated by all subroutines. [1]
Local variables are created inside a subroutine and cannot be accessed directly elsewhere. [1] Local variables
are removed from memory when the subroutine ends. [1] Values can be passed to subroutines as parameters by
value or by reference. [1] Global variables can make a prototype easier to write because values do not need to be
passed repeatedly. [1] However, global variables stay in memory for the whole program and may become
memory-intensive as a program grows. [1] They can also create coupling, testing, debugging and maintenance
problems. [1] A larger program is usually better designed with local variables and parameters. [1]

Page 30
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

2025 - H446/02 Paper 2

Question:
Describe the steps a bubble sort takes to sort items into ascending order.

Answer:
A bubble sort compares adjacent pairs of items. [1] If the pair is already in the correct order, no swap is made. [1]
If the pair is in the wrong order, the items are swapped. [1] This comparison continues to the end of a pass. [1]
Passes are repeated until the list is sorted or the maximum number of passes has been reached. [1]

Question:
Explain how a bubble sort can be made more efficient.

Answer:
A swap flag can be used to record whether a swap has occurred in a pass. [1] If no swaps are made in a pass, the
algorithm can stop because the data is already sorted. [1]

Question:
Identify the type of iteration used by a for loop.

Answer:
A for loop is count-controlled or definite iteration. [1]

Question:
Describe the role of the partition function in quicksort.

Answer:
The partition function selects a pivot. [1] It compares each value with the pivot. [1] Values less than the pivot are
placed on one side and values greater than the pivot are placed on the other side. [1]

Question:
Describe what divide-and-conquer means.

Answer:
Divide-and-conquer breaks a problem into smaller sub-problems. [1] Each smaller sub-problem is solved
separately. [1] The smaller solutions are combined to produce the overall solution. [1]

Question:
Compare linear search and binary search for a large sorted file.

Answer:
A linear search checks items sequentially from the first item onwards. [1] It does not require the data to be sorted.
[1] Its average and worst-case time complexity is O(n), so search time grows linearly with the number of records.
[1] A binary search requires sorted data. [1] It repeatedly compares the target with the middle item and discards
half of the remaining search space. [1] Its average and worst-case time complexity is O(log n), so it scales much
better for a very large file. [1] Both searches can use O(1) extra space. [1] For a large sorted file, binary search
should be used because it is much more time-efficient. [1]

Page 31
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Describe what data mining means.

Answer:
Data mining analyses large quantities of data. [1] It converts data into useful information. [1] It identifies patterns,
trends, anomalies or relationships that may not be obvious to the user. [1]

Question:
Explain how data mining can improve a website and customer experience.

Answer:
Data mining can identify patterns in searches, views or purchases. [1] The company can use these patterns to
personalise adverts, recommend products, promote relevant features or stock more popular items. [1]

Question:
Discuss the use of pipelining by a processor.

Answer:
Pipelining divides instruction processing into stages such as fetch, decode and execute. [1] Different instructions
can be in different stages at the same time. [1] For example, one instruction can be fetched while another is
decoded and another is executed. [1] This reduces processor idle time and can increase instructions completed
per second. [1] It is especially useful when millions of instructions must be processed. [1] However, branch or
jump instructions may require the pipeline to be reset, reducing efficiency. [1]

Question:
Explain why a graph is a visualisation of a problem.

Answer:
A graph is a graphical representation of the problem. [1] It uses nodes to represent entities such as locations and
edges to represent links or routes between them. [1] This simplifies the problem and makes it easier to
understand. [1]

Question:
Describe how A* uses heuristics efficiently.

Answer:
A heuristic estimates the remaining cost or distance. [1] A* adds the heuristic to the distance already travelled to
choose the most promising next node. [1] It avoids exploring every possible route, which can speed up the
search. [1]

Question:
Identify and describe IDE features used for debugging.

Answer:
Breakpoints stop the program at set positions so the flow or variable contents can be checked. [1] Stepping runs
the code one line at a time so the programmer can trace execution. [1] A variable watch window displays variable
or data-structure contents while the program runs. [1] Error diagnostics locate or report details about errors. [1]

Page 32
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Explain the difference between global and local variables.

Answer:
A global variable is accessible throughout the program. [1] A local variable exists only within the scope in which it
is declared, such as a function or procedure. [1]

Question:
Describe drawbacks of using global variables.

Answer:
A global variable exists throughout the run-time of the program, which can waste memory. [1] It can make the
program harder to maintain because it may be difficult to see where the variable is changed. [1] A change in one
part of the program can affect another part, reducing data integrity. [1] Global variables can also reduce
modularity because modules are no longer fully self-contained. [1]

Question:
Describe drawbacks of using local variables.

Answer:
A local variable can only be accessed inside the subroutine where it is declared. [1] Data may need to be passed
as parameters or returned from functions. [1] This can make programming more complex or time-consuming. [1]

Question:
Describe what parameter passing by value means.

Answer:
The function receives a copy of the variable. [1] Changes are made to the copy, not the original value. [1] The
copy is deleted or becomes unavailable when the function ends. [1]

Question:
Describe the drawback of an unbalanced binary search tree.

Answer:
An unbalanced binary search tree can take longer to search. [1] In the worst case, searching may become O(n)
rather than O(log n). [1] More levels or iterations may need to be processed before the value is found. [1]

Question:
Describe features of a linked list data structure.

Answer:
A linked list has a head pointer that points to the first node. [1] Each node contains a data item and a pointer to
the next node. [1] The last node has a null pointer to indicate the end of the list. [1] Data must be accessed by
following pointers rather than by direct indexing. [1]

Page 33
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

Question:
Identify computational methods that can be used when designing a program.

Answer:
Problem recognition identifies what problem needs to be solved. [1] Decomposition splits the problem into smaller
sub-problems. [1] Abstraction removes unnecessary detail. [1] Other valid methods include backtracking,
heuristics, performance modelling, visualisation, algorithmic thinking and logical reasoning. [1]

Question:
Compare using object-oriented classes and a two-dimensional array to implement a linked list.

Answer:
An object-oriented approach can create a new node object whenever one is needed. [1] It does not require the
maximum number of nodes to be known in advance. [1] Classes can be reused in other programs with only small
changes. [1] Encapsulation can keep node attributes private and accessed through methods. [1] A
two-dimensional array implementation requires a fixed maximum size. [1] Unused array elements can waste
memory, but a free list can track where new nodes can be inserted. [1] The array method may be less flexible and
requires more pointer-management code. [1] The object-oriented method is usually more reusable and
memory-efficient for a general linked-list library. [1]

Page 34
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

High-Frequency Mark Scheme Method

Use these patterns whenever a question asks you to define, describe, explain, compare or evaluate Paper 2
theory. Build answers from precise definitions plus one or two consequences.

Decomposition
Define it as breaking a problem into smaller sub-problems. [1] Add that each part can be solved, coded or tested
separately. [1] Link to reuse, team working, easier debugging or simpler design. [1]

Abstraction
Define it as removing unnecessary detail. [1] State that only essential features are kept. [1] Link to reduced
complexity, reduced memory/processing requirements and focus on the core problem. [1]

Problem recognition
State that the first task is identifying exactly what problem must be solved. [1] Mention inputs, outputs, constraints
and whether a computational solution is possible. [1]

Algorithms and Big O

Big O describes how time or space grows as input size increases. [1] Use named forms: O(1) constant, O(log n)
logarithmic, O(n) linear, O(n^2) quadratic, O(2^n) exponential. [1] Always link the complexity to scalability, not just
speed. [1]

Searching and sorting

For search theory, state the precondition for binary search: sorted data. [1] For sort comparisons, compare both
time complexity and space complexity. [1] Mention that the same Big O does not mean identical real running time.
[1]

Stacks and queues

A stack is last in, first out. [1] A queue is first in, first out. [1] Always name the pointer involved: stack pointer, head
pointer or tail pointer. [1] For operations, mention underflow or overflow checks when relevant. [1]

Trees and graphs

A tree has a root, hierarchy and no cycles. [1] A graph uses nodes and edges and may have cycles, direction or
weights. [1] For traversal, breadth-first works level by level; depth-first follows a branch and backtracks. [1]

Functions, procedures and parameters

A function returns a value. [1] A procedure performs a task without returning a value in that way. [1] Passing by
value sends a copy; passing by reference gives access to the original. [1]

Recursion
A recursive routine calls itself. [1] It must have a base case. [1] Each call uses stack memory for its state, so
recursion can use more memory than iteration. [1]

Local and global variables

Local variables have limited scope and can be removed when the subroutine ends. [1] Global variables are
accessible throughout the program and stay in memory. [1] Global variables can simplify access but increase side
effects, coupling and debugging difficulty. [1]

OOP concepts
A class is a template for objects. [1] An object is an instance of a class. [1] Attributes store data and methods
define behaviour. [1] Encapsulation protects attributes by forcing access through methods. [1]

IDE features
Syntax highlighting helps identify keywords and errors. [1] Breakpoints pause execution at chosen lines. [1]
Stepping runs one line at a time. [1] A watch window shows variable values during execution. [1] Auto-complete

Page 35
Mark Scheme Method - OCR H446/02 Paper 2 Written Q&A

reduces typing mistakes. [1]

Evaluation answers
Give both sides. [1] Apply to the given context. [1] End with a justified judgement, such as which method is more
suitable and why. [1]

Concurrent processing
Define it as multiple processes executing at the same time or appearing to do so. [1] Link it to time slicing, threads
or multiple processors. [1] In server scenarios, mention multiple users and the need to protect shared data. [1]

Data mining
Define it as analysing large data sets to find patterns, trends, anomalies or hidden relationships. [1] Apply it to
user behaviour, recommendations, forecasting or targeted improvements. [1] Mention privacy or misuse concerns
in evaluation answers. [1]

Graph algorithms
For A*, define a heuristic as an estimated remaining cost. [1] State that distance travelled plus heuristic is used to
choose a promising next node. [1] For Dijkstra, emphasise known shortest distance from the start node rather
than a heuristic. [1]

Pipelining
State that instruction processing is split into stages such as fetch, decode and execute. [1] Explain that different
instructions can be in different stages at the same time. [1] Add that this reduces processor idle time, but
branching can reduce efficiency. [1]

End of Paper 2 written-theory booklet.

Page 36