IB Computer Science HL flashcards
IB Computer Science HL · Build recall with a focused review session.
32 cards. Read the term and the definition here. The deck at the top of this page is optional.
- Direct changeover
- Deployment where the old system is replaced by the new one in a single step. It is fast and avoids duplicate running costs but carries the highest risk if the new system fails.
- Parallel running
- Deployment where old and new systems run simultaneously for a period. Outputs can be compared and the old system is a fallback, but data entry doubles and costs rise.
- Data migration
- Transferring existing data into a new system, usually with reformatting and validation. Risks include truncation, loss of precision, orphan records, and duplicate entries.
- Acceptance testing
- Testing performed with the client using real data against agreed criteria to confirm the delivered system meets its requirements before it is formally accepted.
- Fetch-execute cycle
- The repeated process where the CPU loads an instruction from the address in the program counter, decodes it, executes the operation, and stores any result.
- Address bus width
- The number of parallel address lines determines how many distinct memory locations can be referenced; n lines allow 2^n addressable locations.
- Two's complement
- A signed binary representation where the most significant bit indicates sign. To negate, invert all bits and add one; range is -2^(n-1) to 2^(n-1)-1.
- Cache locality
- Temporal locality means recently used data is likely reused soon; spatial locality means nearby addresses are likely accessed. Both raise cache hit rates and effective speed.
- Pipelining hazard
- A condition preventing the next instruction from executing in the intended cycle. Data, control, and structural hazards are mitigated by forwarding, stalling, and branch prediction.
- TCP versus UDP
- TCP is connection-oriented with sequencing, acknowledgements, and retransmission. UDP is connectionless with no delivery guarantee, suited to streaming and simple queries.
- Subnet mask function
- A bit pattern that separates the network portion from the host portion of an IP address; host bits all zero give the network address, all one give the broadcast address.
- DHCP
- A protocol that automatically supplies IP address, subnet mask, default gateway, and DNS server settings to devices joining a network, avoiding manual configuration.
- NAT
- Network Address Translation rewrites private source addresses to a public address at a gateway, letting many internal hosts share one routable address.
- TTL field
- A packet header value decremented at each router and discarded at zero, preventing endless circulation; traceroute varies it to reveal the path.
- Abstraction
- Removing detail irrelevant to the current problem so the essential features can be modelled. It simplifies design and lets different implementations share one interface.
- Binary search precondition
- Binary search repeatedly halves a sorted collection, giving O(log n) comparisons. It requires the data to be ordered and supports efficient access to the middle element.
- Merge sort complexity
- A divide-and-conquer sort that splits the collection, sorts halves recursively, and merges them. It is O(n log n) and stable but needs O(n) extra space.
- Recursion base case
- A condition that stops recursive calls without further recursion. Every recursive algorithm needs one, plus progress toward it, or it overflows the call stack.
- De Morgan's laws
- Boolean identities stating that NOT (A AND B) equals (NOT A) OR (NOT B), and NOT (A OR B) equals (NOT A) AND (NOT B). They simplify logic expressions.
- Stack LIFO
- A last-in-first-out structure with push, pop, and peek at one end. Operations at the top are O(1); searching requires removing or traversing elements.
- Queue FIFO
- A first-in-first-out structure where enqueue adds at the rear and dequeue removes from the front. It models scheduling, buffering, and breadth-first traversal.
- Linked list trade-off
- Insertion or deletion at a known node is O(1) and memory grows dynamically, but indexed access is O(n) because nodes must be traversed sequentially.
- BST in-order traversal
- Visiting a binary search tree's left subtree, then node, then right subtree outputs keys in ascending order; search cost is O(log n) only when balanced.
- Adjacency matrix
- A V by V grid storing whether an edge exists between each vertex pair. Edge lookup is O(1) and memory is O(V^2), wasteful for sparse graphs.
- Hash collision handling
- When two keys map to one bucket, chaining stores a list per bucket and open addressing probes other slots. Load factor triggers resizing to keep operations near O(1).
- Round robin scheduling
- Each ready process receives a fixed time quantum in turn. It is fair and responsive for interactive systems, but many context switches reduce throughput.
- Virtual memory
- A technique mapping logical addresses to physical frames via page tables, letting programs exceed physical RAM; excessive page faults cause thrashing.
- Deadlock conditions
- Four simultaneous conditions: mutual exclusion, hold-and-wait, no preemption, and circular wait. Breaking any one prevents deadlock from occurring.
- Closed-loop control
- Control that measures output with a sensor, compares it to a set point, and adjusts the actuator using the error, improving accuracy and rejecting disturbances.
- PID controller
- A controller combining proportional response to current error, integral action removing steady-state offset, and derivative action damping overshoot, tuned by gain settings.
- ADC aliasing
- If an analogue signal is sampled below twice its highest frequency, the Nyquist rate, higher frequencies appear as false lower ones and cannot be recovered.
- IA success criteria
- Testable statements agreed with the client describing what the finished computational solution must do; they drive design, testing, and the final evaluation.