Software Quality Assurance Questions and Answers

Answer: B

ENVIRONMENT is not one of the eight principles.

The Software Engineering Code of Ethics and Professional Practice, adopted by the ACM and the IEEE Computer Society in 1999, sets out eight principles: PUBLIC, CLIENT AND EMPLOYER, PRODUCT, JUDGMENT, MANAGEMENT, PROFESSION, COLLEAGUES and SELF.

PRODUCT, PUBLIC and PROFESSION are all on the list. Environmental concerns can fall under acting in the public interest, but ENVIRONMENT is not a principle in its own right.

Answer: C

The 'Defined' level is Level 3.

The CMM has five maturity levels: 1 Initial (ad hoc), 2 Repeatable (basic project management), 3 Defined (a documented, standard process used across the organisation), 4 Managed (processes measured quantitatively), and 5 Optimizing (continuous process improvement).

Level 1 has no defined process, Level 2 repeats past project successes but has no organisation-wide standard, and Level 4 adds quantitative measurement on top of the defined process.

Answer: C

The true assertion is Adherence to coding standards is checked during code inspection.

A code inspection is a formal review in which the team checks the code against a list of common errors and the project's coding standards.

The others are false: inspection is normally done before unit testing, not after; inspection (formal, checklist-driven) and walkthrough (informal, led by the author) are different techniques; and a walkthrough is carried out by the development team and peers, not an independent test team.

Answer: A

The key's answer is Cyclomatic complexity.

McCabe's cyclomatic complexity counts the independent paths through a module's control flow. High values mean code that is harder to understand, test and change safely, so it is the metric most directly used to judge maintainability, and modules above about 10 are commonly flagged for refactoring.

Halstead volume measures the size of the program's vocabulary of operators and operands, lines of code measure size only, and function points measure delivered functionality, which is used for estimating effort. Halstead volume and lines of code also feed composite indexes of maintainability, but on their own they say less about how hard code is to change.

Answer: C

KPIs serve for Monitoring progress toward strategic objectives.

A key performance indicator is a measurable value chosen to show how well an organisation or system is achieving its key goals, such as order processing time, customer satisfaction or system availability. In system analysis KPIs define what success looks like and track whether a system is delivering it.

Hardware efficiency, database schemas and test management are narrower technical concerns. A KPI may measure one of them, but the role of KPIs is to tie performance to strategic objectives.

Answer: B

Efficiency does not include licensing.

Efficiency is one of the essential attributes of good software: it should not waste system resources. Sommerville describes it as including responsiveness, processing time and memory utilisation.

Licensing is a legal and commercial matter, how the software may be used and sold, and has nothing to do with how efficiently it runs.

Answer: A

The maturity levels of the Capability Maturity Model used to measure a process are Initial, Repeatable, Defined, Managed and Optimizing. They run from level 1 where the process is ad hoc to level 5 where it is continuously improved.

The names 'Primary' and 'Secondary' and 'Stating' in the other options are not levels of the model, so the second and third lists are wrong, and 'None of the above' is wrong because the first list gives the levels correctly.

Answer: D

Software quality assurance is an umbrella activity applied throughout the software engineering process. It covers all steps from requirements to maintenance, with reviews, standards and measurement to be sure that quality is built in.

Debugging, testing and designing are activities within particular phases: debugging follows the finding of an error, testing is done after the code is written, and designing comes before it. They do not run through the whole process.

Answer: D

The quality of a software system is determined by all the factors named. Correctness and reliability are the first set, efficiency, usability and maintainability the second, and testability, portability, accuracy, error tolerance, expandability, access control and audit the third. Each list is an attribute of quality in the classical quality models, so no single option is complete alone. The keyed answer is all of the above.

Answer: D

The program P fails if either of the two subprograms fails. The probability that P1 does not fail is 0.5 and the probability that P2 does not fail is 0.6, so the probability that neither fails is

and the program fails in the remaining 1 - 0.3 = 0.7 of the runs, that is 70 per cent, taking the failures to be independent. The figure 50 per cent is the failure rate of P1 alone, 60 per cent has no basis in the data, and 10 per cent is the difference of the rates. The keyed answer is 70%.