Algorithms, Flowcharts and SDLC: Symbols, Models and PYQs

The correct answer is C, diamond. The diamond or rhombus is the decision symbol: it holds a condition and has one entry with two exits, marked Yes and No, so the flow branches according to the test. Option A is wrong because the oval, also drawn as a rounded rectangle, is the terminal symbol used only for Start and Stop. Option B is wrong because the rectangle is the process symbol, used for a calculation or an assignment such as adding two numbers and storing the sum. Option D is wrong because the parallelogram is the input or output symbol, used when a value is read from the user or a result is printed. Candidates most often swap the rectangle and the parallelogram, so fix the pair in memory: rectangle does the work, parallelogram talks to the outside world, and the diamond decides.

The correct answer is A, Start and Stop. The oval is the terminal symbol, and a well-formed flowchart has exactly one Start terminal and at least one Stop terminal; a chart that can never reach a Stop describes an infinite loop and so breaks the finiteness property of an algorithm. Option B is wrong because a subroutine or module call is shown by the predefined process symbol, a rectangle with double lines along its sides. Option C is wrong because a decision is shown by the diamond, which branches into Yes and No. Option D is wrong because the direction of flow is shown by the arrow or flow line, which carries the eye from one symbol to the next; by convention the chart is read from top to bottom and from left to right. Learning the seven standard symbols covers almost every flowchart question.

The correct answer is B, Al-Khwarizmi. Muhammad ibn Musa al-Khwarizmi was a ninth-century Persian mathematician and astronomer whose treatise on calculation with Hindu-Arabic numerals reached Europe in Latin translation; his Latinised name produced the word algorithm, and another of his works gave the word algebra. Option A is wrong because Euclid is remembered for geometry and for the Euclidean algorithm for the greatest common divisor, but the word itself does not come from his name. Option C is wrong because Charles Babbage designed the Difference Engine and the Analytical Engine and is called the father of the computer, not the source of this word. Option D is wrong because Blaise Pascal built the Pascaline, an early mechanical adding machine, and has a programming language named after him. Expect this as a one-mark definition question.

The correct answer is C, O(log n). Binary search compares the key with the middle element and throws away half the array at every step, so the number of comparisons grows with the logarithm of the array size; an array of a million sorted items needs about twenty comparisons. The method works only on data that is already sorted, which is the condition examiners like to test. Option A is wrong because O(n) is linear search, which looks at the elements one after another and does not need sorted data. Option B is wrong because O(n log n) is the complexity of the efficient sorting algorithms such as merge sort and heap sort, not of a search. Option D is wrong because O(n squared) describes the simple sorts, namely bubble sort, selection sort and insertion sort in their worst case. Remember that Big-O keeps only the dominant term.

The correct answer is C, the Waterfall model. The Waterfall model is called the linear sequential model because the project flows in one direction through requirement analysis, design, coding, testing, deployment and maintenance, with each phase completed and signed off before the next begins and no overlap between them. It suits only projects whose requirements are frozen at the start. Option A is wrong because the spiral model, given by Barry Boehm, is risk-driven and repeats planning, risk analysis, engineering and evaluation in loops, so it is anything but linear. Option B is wrong because the prototype model builds a throwaway working mock-up to collect user feedback before the real build. Option D is wrong because RAD, or rapid application development, assembles the product quickly from reusable components with heavy user involvement, again not a single straight line of phases.

The correct answer is A, requirement gathering and analysis. This is the first phase of the SDLC: the team meets the users, records what the system must do and what constraints apply, and sets it all down in the Software Requirements Specification, which then becomes the agreed basis for everything that follows and the yardstick for acceptance testing. Option B is wrong because the design phase converts the already approved SRS into a high-level design of modules and a low-level design of their logic. Option C is wrong because coding writes the program from the design documents, and by then the requirements are supposed to be settled. Option D is wrong because maintenance is the last and usually the longest and costliest phase, covering corrective, adaptive and enhancement work after delivery. A defect caught in the requirement phase is the cheapest to repair.

The correct answer is A: coupling measures the degree of functional interdependence between two modules. If one module must know the internal data or logic of another to work, the two are tightly coupled, and a change in one forces a change in the other. Good design therefore aims at low coupling together with high cohesion, where cohesion measures how closely the parts inside a single module belong together. Option B is wrong because it states the opposite, non-dependence; that idea belongs to the goal of loose coupling, not to the definition of coupling itself. Option C is wrong because robustness is the ability of a module to behave sensibly on wrong or extreme input, which is a quality attribute tested during testing, not a measure of inter-module linkage. Option D is wrong because it is a loosely worded restatement that misses the word interdependence on which the standard definition rests.

The correct answer is B, Barry Boehm. Boehm described the spiral model, in which the project moves outward through repeated loops and each loop performs four activities: determining objectives, risk analysis and resolution, development and testing, and planning the next loop. Because risk is examined before every stage of build, the model suits large and uncertain projects. Option A is wrong because Winston Royce is associated with the paper that set out the stage-wise waterfall approach, the linear sequential model. Option C is wrong because Edsger Dijkstra is remembered for the shortest path algorithm, for structured programming and for his criticism of the goto statement, not for a life cycle model. Option D is wrong because Grady Booch is one of the three authors of the Unified Modeling Language and of the Rational Unified Process, which is related but distinct from the spiral model.

The correct answer is B, only I. A code walk-through is an informal review technique: a few team members read a module line by line, mentally execute it on sample data and note the defects they find, without the checklists, metrics and formal records of an inspection. Statement I is therefore correct. Statement II is wrong on timing, because a walk-through can only be held after the module has been coded; there is no code to walk through before coding. That makes option D, only II, wrong, and option C, both, wrong as well. Option A is wrong because statement I is sound. The learning point is the pair: a walk-through is informal and conducted after coding, while a code inspection is formal, uses a checklist of common error types and produces a written report.

The correct answer is B, black-box testing. In black-box testing the tester treats the program as a sealed box, supplies inputs drawn from the specification and compares the outputs with what the specification promises; equivalence partitioning and boundary value analysis are its standard techniques, and no knowledge of the source code is needed. Option A is wrong because white-box or glass-box testing is the opposite: the tester reads the code and designs cases to cover statements, branches and paths. Option C is wrong because unit testing names the level at which a single module is tested, not the method of designing the cases, and a unit test may itself be black-box or white-box. Option D is wrong because regression testing means re-running old tests after a change, to confirm that the change has not broken what already worked. Level and technique are different axes.