Implementation of Boundary Value Analysis

Study Case Implementation of Boundary Value Analysis in Black Box Testing-Based Functional Testing Michael Kurniawan (2018123456) University of Cincinnati, 2600 Clifton Ave, Cincinnati, OH 45221, United States kurniaml@mail.uc.edu Abstract Software testing is the process of running a program with the aim of identifying errors that may not have been detected previously. Effective test cases are those that have a high probability of finding hidden defects, while testing is considered successful if it is able to reveal errors that were not detected in the previous stage. This study applies the Black Box Testing technique, specifically the Boundary Value Analysis method, which is a testing approach that focuses on evaluating the upper and lower boundary values of the tested input. This method is applied to the class addition function in the College Admission Prediction Application, which is then adapted to the United States context. The test results show that the data validation process still has a number of shortcomings, which could potentially result in data storage that is inconsistent with the expected values in the database. These findings can be used as a basis for optimizing the overall quality of the application. Keywords: Testing, Black Box Testing, Boundary Value Analysis. INTRODUCTION The National Selection for State University Admission Prediction Application was adapted into the College Admission Prediction Application, which is an application designed to assist school counselors in guiding students to choose public universities in the United States that match their academic performance. This application is expected to support students' success in the college admission process. The application has a number of modules, including class management, student management, grade management, university and major data management, ranking based MICHAEL KURNIAWAN 1 Study Case on grades, and other relevant functional modules. After the development process is complete, the application needs to be tested to ensure that all functions work according to user requirements. Testing is the process of running a program with the aim of finding previously unidentified errors. A good test case is one that is able to reveal hidden defects, while a successful test is one that manages to find new errors. The main objective of testing is to design scenarios that can systematically reveal errors with minimal use of time and resources. LITERATURE REVIEW Software Testing Software testing is very important to ensure that applications work as designed. Developers and testers must provide special sessions to evaluate the software so that errors can be found early and corrected immediately. Testing is a key element in software quality assurance (SQA) and an integral part of the software development lifecycle, such as analysis, design, and coding [1]. Various testing strategies have been developed, each providing a framework for developers. According to Johnson and Carter [2], a good testing strategy has the following characteristics: 1. Testing begins at the module level before moving on to overall system integration. 2. Different testing techniques are used at different phases in the development process. 3. Testing is performed by developers, and for large projects, also by an independent testing team. 4. Testing and debugging are two different processes, but debugging must be accommodated in every testing strategy. Software testing is part of a broader process, namely Verification and Validation (V&V), which includes: 1. Verification, ensuring that the software implements the specified functions. 2. Validation, ensuring that the software meets customer requirements. V&V activities include technical reviews, configuration audits, and system performance monitoring [2]. MICHAEL KURNIAWAN 2 Study Case There are several types of software testing, including: 1. Software engineering methods as the basis for quality control. 2. Analysis, design, and construction methods to produce consistent quality. 3. Formal Technical Reviews to ensure product quality at every stage. 4. Quality measurement of each element of the software configuration. 5. Standards and procedures that ensure consistency in the SQA process. 6. Testing methods as the final stage to ensure that the software is free of major errors. Testing principles according to Davidson [3]: 1. All tests must be traceable to customer requirements. 2. Test plans must be created before the testing process begins. The Pareto principle applies: 80% of errors are found in 20% of modules. 3. Testing is performed starting with small modules before moving on to larger systems. 4. Comprehensive testing is impossible, so effective testing is performed by an independent party. The testing stages consist of: 1. Unit Testing – testing specific modules to maximize error detection. 2. Integration Testing – ensuring that each unit works well when combined. 3. High-Order Testing – performed after the entire system is integrated. 4. Validation Testing – ensuring that the software meets functional and performance requirements. Gilbert [2] provides procedures for implementing effective testing strategies, including setting quantitative requirements, understanding user profiles, implementing rapid cycle testing, and conducting continuous evaluation. Types of software testing according to Harris [4]: 1. White Box Testing, which is testing based on the details of the internal structure and logic of the program. 2. Black Box Testing, which is testing based on functional specifications without knowing the code structure. White Box Testing White Box Testing is a software testing method that evaluates internal logic, code structure, and program control flow [5]. If the output does not match expectations, the module needs to be recompiled and evaluated until the error can be corrected. MICHAEL KURNIAWAN 3 Study Case The general stages of White Box Testing include: 1. Testing all logical decisions. 2. Testing all loops according to the limitations. 3. Testing the internal data structure and its validity. Advantages of White Box Testing [5]: 1. Detection of logical errors, including errors in if or loop structures. 2. Identification of assumption inconsistencies, so that initial estimates can be corrected. 3. Detection of typing errors, especially in case-sensitive programming languages. The disadvantage is that White Box Testing requires a lot of resources and time when applied to large-scale software. Black Box Testing Black Box Testing focuses on evaluating the functional specifications of software without looking at the internal structure of the code. In this method, testers define various input conditions and assess whether the resulting output matches the specified requirements. This technique is not an alternative to White Box Testing, but rather serves as a complement to ensure that aspects not covered by White Box Testing are still properly tested. Generally, Black Box Testing can identify: 1. Functions that do not work properly or are unavailable. 2. Interface errors. 3. Errors in data structures and database access. 4. Performance errors. 5. Errors in the initialization and termination processes. This testing aims to answer several important questions, including: 1. How can functions be tested to be declared valid? 2. What types of input can be used as good test cases? 3. Is the system sensitive to certain input combinations? 4. How can a set of data be isolated for testing purposes? 5. What is the average amount of data that can be processed by the system? 6. What effects occur when certain data combinations are given in system operations? Some common techniques in Black Box Testing include: 1. Equivalence Partitioning MICHAEL KURNIAWAN 4 Study Case 2. Boundary Value Analysis / Limit Testing 3. Comparison Testing 4. Sample Testing 5. Robustness Testing 6. Behavior Testing 7. Requirement Testing 8. Performance Testing 9. Endurance Testing 10. Cause-Effect Relationship Testing Boundary Value Analysis / Limit Testing Boundary Value Analysis (BVA) is a Black Box Testing technique that focuses on testing inputs at minimum and maximum limits. The basic principles of BVA include: 1. Many errors are found in input values that are exactly at the limits of the value range. 2. BVA allows the selection of test cases that evaluate input limit values. 3. BVA complements Equivalence Partitioning because it selects elements on the boundary side of each equivalence class. 4. Examples of boundary testing: a. If a value is limited to a range from a to b, then the values tested are (a−1), a, (a+1), (b−1), b, and (b+1). b. If the input requires n values, then (n−1), n, and (n+1) are tested. c. The same rules can be applied to system outputs. d. If the application uses internal data structures such as array bounds or buffer sizes, then testing is performed using values that push those bounds. BVA implementation is generally done by maintaining one variable at its nominal (average) condition, and testing the other variables at their extreme values. These extreme values include: • Min — minimum value • Min+ — value just above the minimum • Nom — average value • Max- — value just below the maximum • Max — maximum value Case example: If the date field consists of: MICHAEL KURNIAWAN 5 Study Case 1 ≤ date ≤ 31 1 ≤ month ≤ 12 1812 ≤ year ≤ 2016 Then every entry outside these limits will generate an error message such as “The date you entered is invalid”. BOUNDARY VALUE ANALYSIS TESTING Based on the previous concept, the BVA method was applied to test the College Admission Prediction Application, specifically the “Add Class” functionality. This feature contains a data entry panel as shown in Figure 1. The form has three input fields, namely major, class name, and academic year. Figure 1. Add Class Form The data entered in the form is stored in a table as shown in Table 1. Table 1. Structure of the “Class” Table The testing then focused on two fields, namely class name and academic year, because the major field is a dropdown with predefined values. MICHAEL KURNIAWAN 6 Study Case The following are the test scenarios and their results: Testing the “Class Name” Field Rule A.1 — Must consist of 3 words Table 2. Test Results for Rule A.1 Rule A.2 — The first word must be the number 12 Table 3. Test Results for Rule A.2 Rule A.3 — The second word must be a major (IPA/IPS) Table 4. Test Results for Rule A.3 Rule A.4 — The third word must be a number between 1 and 10. Table 5. Test Results for Rule A.4 MICHAEL KURNIAWAN 7 Study Case Testing the “Academic Year” Field Rule B.1 — Must consist of 4 digits. Table 6. Test Results for Rule B.1 Rule B.2 — The academic year must be between 2011 and 2015. Table 7. Results of Rule B.2 Test Based on these tests, the minimum total test data that must be prepared is: 2 fields × 2 rules × 3 samples = 12 test data. The test results show that input validation in the application is still incomplete, so the application still accepts data that does not match the expected format. CONCLUSION Based on testing the Add Class feature using the Black Box Testing method with Boundary Value Analysis, the following conclusions were obtained: 1. The Black Box Testing method is easy to apply because it only requires determining the lower and upper limits of the input being tested. 2. The estimated number of test data can be calculated based on the number of input fields, the number of entry rules, and the number of limits evaluated. 3. The test results show that the application still accepts invalid data, which can cause inconsistencies in the stored data. 4. Based on the evaluation using the BVA method, the data entry function needs to be equipped with additional validation to ensure data accuracy and consistency in accordance with functional requirements. MICHAEL KURNIAWAN 8 Study Case REFERENCE MICHAEL KURNIAWAN 9