Sonali Bank Limited Assistant Engineer (IT) Preliminary Exam: 2016প্রশ্ন ও সমাধান

Correct answer: (d) Accumulator

In the classic accumulator-based CPU (for example the Intel 8085), the accumulator, register A, is wired to the ALU. One operand comes from the accumulator, and the ALU writes the result straight back into it, updating the flags. For example, ADD B performs A ← A + B. The program can then move the value from A to another register or to memory with a separate instruction.

Why the other options are wrong

• (a) Memory Address registers: the MAR holds the address of the memory location being accessed, not data results.
• (b) PC: the Program Counter holds the address of the next instruction to fetch.
• (c) General registers: a result can be copied there later, and register-register CPUs can target them, but the immediate, default destination of the ALU in the accumulator model is the accumulator.

Tip: "Accumulator" literally means the register where results accumulate; in 8085 arithmetic/logic instructions the result lands in A.

Source: 3.2 Accumulator (NPTEL)

Correct answer: (c) Stacks

Push and pop are the defining operations of a stack: push places an item on the top and pop removes the top (most recently added) item, giving LIFO order.

Why the other options are wrong

• (a) Array: An array is accessed by index at any position. It can be used to implement a stack, but push and pop are the stack's operations, not the array's.
• (b) Lists: A general list allows insertion and deletion at any position. When a list offers push/pop methods (for example, a Python list's append/pop), it is being used as a stack.
• (d) All of the above: Only the stack is defined by push and pop, so not all three apply.

Exam note: Some languages give arrays/lists methods named push and pop, but in data-structure terms these names belong to the stack; that is the sense the paper intends.

Source: stack — NIST Dictionary of Algorithms and Data Structures

Correct answer: (d) None of them

Real-time systems (industrial control, avionics, medical devices) must respond within a guaranteed time. That needs a real-time operating system (RTOS), whose scheduling makes the completion time of a task predictable. Microsoft explains the difference this way: an ordinary OS gives deterministic results but not a deterministic finishing time, whereas a real-time OS makes the timing deterministic too; a hard real-time OS meets an exact deadline every time. Microsoft's own device-class real-time product was a separate embedded OS (Windows Embedded Compact), not desktop Windows.

Why the other options are wrong

• (a) Windows: general-purpose Windows is tuned for interactive use and throughput; it is not the recommended RTOS.
• (b) Unix: standard Unix is a time-sharing OS that shares the CPU fairly among users, without deadline guarantees.
• (c) Oracle: Oracle Database is a relational database management system, not an operating system.

Exam note: Some vendors sell real-time Unix-like systems, but the question names the general-purpose families, so the expected answer is "None of them".

Tip: RTOS = guaranteed deadline. Windows, Unix and Oracle are the classic wrong answers.

Source: Soft Real-Time (Microsoft Learn); Windows Embedded Compact 2013 (Microsoft Learn); Introduction to Oracle Database (Oracle Help Center)

Correct answer: (c) O(n²)

Bubble sort compares each adjacent pair and swaps it if it is out of order, repeating passes over the list. In the worst case (reverse-sorted input) pass 1 makes n−1 comparisons, pass 2 makes n−2, …, down to 1:
(n−1) + (n−2) + … + 1 = n(n−1)/2 comparisons, e.g. 4,950 for n = 100. That is O(n²) for arbitrary data.

Why the other options are wrong

• (a) O(n): only the best case of the early-exit version on an already sorted list (n−1 comparisons, one pass); not the algorithm's complexity in general.
• (b) O(log n): the cost of binary search; a sort must at least look at all n items.
• (d) O(n log n): the cost of efficient sorts such as merge sort and heapsort, not bubble sort.

Tip: Bubble, selection and insertion sort are O(n²); merge sort and heapsort are O(n log n).

Source: bubble sort (NIST Dictionary of Algorithms and Data Structures)

Correct answer: (d) All of the above

Optical fibre's high bandwidth, low loss and immunity to interference make it useful at every scale:

• LAN in an office: Ethernet has fibre versions (10BASE-F onward up to today's multi-gigabit links), widely used for building backbones and links between floors.
• Telephone lines: Telephone networks have largely replaced their coaxial trunks with fibre, and fibre-to-the-home now carries voice as well as data.
• Data connectivity among branch offices: Fibre WAN and metro links connect offices in different locations at high speed over long distances.

Why the other options are wrong

• (a) LAN in an office: True, but fibre is not limited to LANs.
• (b) telephone lines: True, but only one of several uses.
• (c) data connectivity among the branches of offices: True, but again only one of the uses listed.

Tip: When every listed use of fibre is genuine, "All of the above" is the answer.

Sources: Transmission Media, lecture notes (University of Baghdad); SECL 1.0 Glossary (Cisco)

Correct answer: (d) new

In Java an object is created with the new operator, for example Box b = new Box();. new allocates memory for the object, runs the constructor named after it, and returns a reference to the new object. Since new is not among options (a) to (c), the answer is (d).

Why the other options are wrong

• (a) class: A keyword used to declare (define) a class, not to create an instance of it.
• (b) scanf: A C library input function; it does not exist in Java (Java reads input with classes such as Scanner).
• (c) print: A method of System.out that writes text to the console; it creates nothing.

Tip: Object creation in Java = new + a constructor call, e.g. new ClassName().

Source: Creating Objects (The Java Tutorials)

Correct answer: (a) creating and processing forms

A database system has two layers. The DBMS organises, stores and searches the data: it creates databases and tables, reads and modifies data, enforces integrity rules, controls concurrent access, applies security and handles backup and recovery. The database application sits on top: it builds the screens users work with (data-entry forms), produces reports and runs business logic, sending SQL to the DBMS. Forms are designed around the business's own paperwork and have nothing to do with how the data is stored, so they belong to the application.

Why the other options are wrong

• (b) creating databases: A core DBMS function (CREATE DATABASE, CREATE TABLE and the supporting structures).
• (c) processing data: Reading, inserting, updating and deleting data is the DBMS's main job.
• (d) administrating databases: Security, user permissions, backup and recovery, and concurrency control are DBMS functions.

Tip: DBMS = store, retrieve, protect the data; application = forms, reports and business logic.

Source: Components of a Database System - CSCI 313 notes, Edinboro University