All Job Interview Questions and Answers << Job Questions and Answers for Software Testing - Testing Techniques
Questions and answers taken from real job interviews.
Find interview questions and answers on this website:
Orthogonal array is a two-dimensional array in which if we choose any two columns in the array and all the combinations of numbers will appear in those columns. The following figure shows a simple L9(34) orthogonal array. In this the number 9 indicates that it has 9 rows. The number 4 indicates that it has 4 columns and 3 indicates that each cell contains a 1, 2, and 3. Choose any two columns. Let's choose column 1 and 2. It has (1,1), (1,2), (1,3), (2,1), (2,2), (2,3), (3,1), (3,2), (3,3) combination values. As you can see these values cover all the values in the array. Compare the values with the combination of column 3 and 4 and they will fall in some pair. This is applied in software testing which helps us eliminate duplicate test cases.
In some projects there are scenarios where we need to do boundary value testing. For instance, let's say for a bank application you can withdraw a maximum of 25000 and a minimum of 100. So in boundary value testing we only test the exact boundaries rather than hitting in the middle. That means we only test above the max and below the max. This covers all scenarios. The following figure shows the boundary value testing for the bank application which we just described. TC1 and TC2 are sufficient to test all conditions for the bank. TC3 and TC4 are just duplicate/redundant test cases which really do not add any value to the testing. So by applying proper boundary value fundamentals we can avoid duplicate test cases, which do not add value to the testing.
As the name suggests they are tables that list all possible inputs and all possible outputs. A general form of decision table is shown in the following figure. Condition 1 through Condition N indicates various input conditions. Action 1 through Condition N are actions that should be taken depending on various input combinations. Each rule defines unique combinations of conditions that result in actions associated with that rule.
In equivalence partitioning we identify inputs which are treated by the system in the same way and produce the same results. You can see from the following figure applications TC1 and TC2 give the same results (i.e., TC3 and TC4 both give the same result, Result2). In short, we have two redundant test cases. By applying equivalence partitioning we minimize the redundant test cases. So apply the test below to see if it forms an equivalence class or not: • All the test cases should test the same thing. • They should produce the same results. • If one test case catches a bug, then the other should also catch it. • If one of them does not catch the defect, then the other should not catch it.
Exploratory testing is also called adhoc testing, but in reality it's not completely adhoc. Ad hoc testing is an unplanned, unstructured, may be even an impulsive journey through the system with the intent of finding bugs. Exploratory testing is simultaneous learning, test design, and test execution. In other words, exploratory testing is any testing done to the extent that the tester proactively controls the design of the tests as those tests are performed and uses information gained while testing to design better tests. Exploratory testers are not merely keying in random data, but rather testing areas that their experience (or imagination) tells them are important and then going where those tests take them.
Random testing is sometimes called monkey testing. In Random testing, data is generated randomly often using a tool. For instance, the following figure shows how randomly-generated data is sent to the system. This data is generated either using a tool or some automated mechanism. With this randomly generated input the system is then tested and results are observed accordingly. Random testing has the following weakness: • They are not realistic. • Many of the tests are redundant and unrealistic. • You will spend more time analyzing results. • You cannot recreate the test if you do not record what data was used for testing. This kind of testing is really of no use and is normally performed by newcomers. Its best use is to see if the system will hold up under adverse effects.
There are four types of severity ratings as shown in the table: • Severity 1 (showstoppers): These kinds of defects do not allow the application to move ahead. So they are also called showstopper defects. • Severity 2 (application continues with severe defects): Application continues working with these types of defects, but they can have high implications, later, which can be more difficult to remove. • Severity 3 (application continues with unexpected results): In this scenario the application continues but with unexpected results. • Severity 4 (suggestions): Defects with these severities are suggestions given by the customer to make the application better. These kinds of defects have the least priority and are considered at the end of the project or during the maintenance stage of the project.
As the name specifies semi-random testing is nothing but controlling random testing and removing redundant test cases. So what we do is perform random test cases and equivalence partitioning to those test cases, which in turn removes redundant test cases, thus giving us semi-random test cases.
A negative test is when you put in an invalid input and receive errors. A positive test is when you put in a valid input and expect some action to be completed in accordance with the specification.