Reliability characteristic of a software is defined as a failure free software for operations for a period of time in a particular environment. Reliability testing is performed to make sure that the software is reliable and satisfies the purpose of its development. It gives a stable outcome or results for the steps performed consistently.
Now a days, since the dependency of software is high, people blindly believe the in the software and will trust on its result whatever it may be, users believe that the data shown is correct and always the software operates correctly. The software applications are used everywhere nowadays almost all fields like healthcare, government sectors and telecommunication. The purpose of reliability testing is to ensure that the software provides stable outcome and perfectly satisfies the requirements for which it is made and also rendering the fault-free operation.
Things that are influencing software reliability are:
- The number of bugs or mistakes present in the software.
- The user’s way of operating the system.
Reliability testing helps to identify many problems while designing the software and also verifies the functionality. It is one of the not functional quality characteristic of software. The goal of reliability testing is to check whether the software meets the requirements of client’s reliability. There are various levels of performing this testing. Most complex system will be tested at unit, assembly and subsystem levels.
Factors to measure the reliability software:
- Test – retest reliability:
We will consider a situation in which we will be testing software functionality twice at 10.00 am and again at 12.00 pm later. When we compare the results, suppose if we get high correlation in results we can say that the software is reliable. Consider an example of a person attending an IQ test and scoring 150 points and after some months he takes again IQ tests and scores 60 points than he may be not considered as reliable. Reliability is a consistency of software to provide the stable outcome.
- Parallel form of reliability:
Here many testers, who will conduct tests in many forms at the same time. They verify results in parallel way.
- Inter-related reliability:
It is also called as inter-coder reliability. It is used for special purpose to check the reliability which consists the multiple raters or judges. For example suppose a contestant taking part in a halloween competition and getting marks as 9,8,9 from multiple judges. This score can be considered as reliable. But instead if he had scored 9,3,6 then it is not reliable. Similarly for the software, consistent quality rating are to be provided by all testers.
Reliability and validity are the two important features that a any test scores can have, they are always mentioned together but often give two different kinds of information.
Reliability tells us how consistently the test scores measure something whereas the validity tells us whether the test scores are measuring the right things for a particular use of test.
Consider an analogy which shows the difference:
By testing the measure of test taker’s proficiency in a particular set of knowledge or skills in a shooting target with each shot, which represents the administration of the test.
A test can have a high reliability in one group of test takers than another group .But validity depends on how the scores are being used its reliability will not.