How to Score Higher in OCR A Level Computer Science (H446)

How to fix this

For any pseudocode algorithm question: (1) use // or DIV/MOD explicitly — never write bare '/' when integer division is needed; (2) always assign calculation results to a named variable before using them (a line like 'denaryNum MOD 8' earns no marks if the result is discarded); (3) generalise the solution to handle input of any length, not just a fixed number of digits; (4) ensure the output order is correct — if building a result by appending remainders, either append in reverse or reverse the final string.

How to fix this

Practise writing complete class definitions end-to-end: constructor (__init__ in Python) sets private attributes using the parameter values passed in (not the other way round). Getter methods must be declared as functions (not procedures) and must return the private attribute value — do not print, do not set. When accessing an object's data in another method, always call the getter method (e.g. obj.getValue()) rather than accessing the attribute directly. Mark the distinction between a class declaration and object instantiation — they are separate steps.

How to fix this

Distinguish polynomial from exponential: O(n²) means the time grows as the square of n (e.g. bubble sort worst case) — this is polynomial. O(2ⁿ) means the time doubles with every additional input item — this is exponential. Never define 'exponential complexity' as 'grows exponentially'; the mark requires a description in terms of n (e.g. 'the number of operations doubles each time n increases by 1'). For logarithmic O(log n): explain that the amount of additional work required becomes progressively smaller as n grows, not just that it is proportional to log n.

How to fix this

The key insight: merge sort splits the data into individual items (base case: a single item is trivially sorted), then rebuilds sorted order by merging pairs of already-sorted sub-lists using two pointers. During the merge phase, the pointers compare the front items of the two sub-lists; the smaller is appended to the output list and that pointer advances. No sorting happens within sub-lists — all ordering is achieved by the merge comparisons. Questions frequently test whether you can describe the merge phase accurately, not just the split phase.

How to fix this

Concurrent processing = multiple processes appear to run simultaneously on a single processor via time-slicing or scheduling; at any instant only one process uses the CPU. Parallel processing = multiple instructions genuinely execute at the same time on multiple processors/cores. Saying 'many things processed at the same time' without specifying what the things are earns no marks. For benefits of concurrency: say 'less CPU idle time' or 'more tasks completed within a given time unit' — not just 'quicker' or 'efficient', which are too vague. Note: pipelining is not concurrent processing.

How to fix this

Know the precise vocabulary for each structure: trees have nodes (not vertices), branches (not edges), and are hierarchical, rooted and acyclic. Graphs have vertices and edges (directed or undirected, weighted or unweighted), may contain cycles and need not be connected. When comparing a tree to a graph, state that a tree is a connected, acyclic subgraph. Binary search tree properties: (1) each node has at most two children; (2) all nodes in the left sub-tree are less than the parent; (3) all nodes in the right sub-tree are greater than or equal to the parent.

How to fix this

The correct pop() sequence is: (1) check the stack is not empty (topStack > 0 or equivalent); (2) store the value at the current top (data[topStack]); (3) decrement the topStack pointer; (4) return the stored value. Returning the value inside the function immediately ends the function — any code after a return statement does not execute. So if you return topStack - 1 first, the decrement on the next line is never reached. Write out each step explicitly and ensure the decrement occurs before the return.

How to fix this

Level 1 = isolated facts or definitions. Level 2 = correct description with some reference to the scenario. Level 3 = evaluative comparison relevant to the specific context. To reach Level 3: take each key point and ask 'so what does this mean for this particular scenario?' For example, when asked about database types for a growing membership system, it is not enough to describe relational and non-relational databases — you must evaluate which is better suited as membership scales, citing specific features such as referential integrity, schema flexibility or query performance. Conclude with a justified recommendation.